IceFloeTracker
Documentation for IceFloeTracker.jl package.
Functions
Base.isequal — Methodisequal(matchedpairs1::MatchedPairs, matchedpairs2::MatchedPairs)Return true if matchedpairs1 and matchedpairs2 are equal, false otherwise.
IceFloeTracker._adjust_histogram — Method_adjust_histogram(masked_view, nbins, rblocks, cblocks, clip)Perform adaptive histogram equalization to a masked image. To be invoked within imsharpen.
Arguments
masked_view: input image in truecolor
See imsharpen for a description of the remaining arguments
IceFloeTracker._bin9todec — Method_bin9todec(v)Get decimal representation of a bit vector v with the leading bit at its leftmost posistion.
Example
julia> _bin9todec([0 0 0 0 0 0 0 0 0])
+IceFloeTracker · IceFloeTracker.jl IceFloeTracker
Documentation for IceFloeTracker.jl package.
Functions
Base.isequal — Methodisequal(matchedpairs1::MatchedPairs, matchedpairs2::MatchedPairs)
Return true if matchedpairs1 and matchedpairs2 are equal, false otherwise.
sourceIceFloeTracker._adjust_histogram — Method_adjust_histogram(masked_view, nbins, rblocks, cblocks, clip)
Perform adaptive histogram equalization to a masked image. To be invoked within imsharpen.
Arguments
masked_view: input image in truecolor
See imsharpen for a description of the remaining arguments
sourceIceFloeTracker._bin9todec — Method_bin9todec(v)
Get decimal representation of a bit vector v with the leading bit at its leftmost posistion.
Example
julia> _bin9todec([0 0 0 0 0 0 0 0 0])
0
julia> _bin9todec([1 1 1 1 1 1 1 1 1])
-511
sourceIceFloeTracker._branch_filter — Function_branch_filter(
+511
sourceIceFloeTracker._branch_filter — Function_branch_filter(
img::AbstractArray{Bool},
func1::Function=branch_candidates_func,
-func2::Function=connected_background_count,)
Filter img with _operator_lut using lut1 and lut2.
sourceIceFloeTracker._operator_lut — Method_operator_lut(I, img, nhood, lut1, lut2)
Look up the neighborhood nhood in lookup tables lut1 and lut2.
Handles cases when the center of nhood is on the edge of img using data in I.
sourceIceFloeTracker.absdiffmeanratio — Methodabsdiffmeanratio(x, y)
Calculate the absolute difference between x and y divided by the mean of x and y.
sourceIceFloeTracker.add_padding — Methodadd_padding(img, style)
Extrapolate the image img according to the style specifications type. Returns the extrapolated image.
Arguments
img: Image to be padded.style: A supported type (such as Pad or Fill) representing the extrapolation style. See the relevant documentation for details.
See also remove_padding
sourceIceFloeTracker.add_passtimes! — Methodadd_passtimes!(props, passtimes)
Add a column passtime to each DataFrame in props containing the time of the image in which the floes were captured.
Arguments
props: array of DataFrames containing floe properties.passtimes: array of DateTime objects containing the time of the image in which the floes were captured.
sourceIceFloeTracker.addfloemasks! — Methodaddfloearrays(props::DataFrame, floeimg::BitMatrix)
Add a column to props called floearray containing the cropped floe masks from floeimg.
sourceIceFloeTracker.addmatch! — Methodaddmatch!(matched_pairs, newmatch)
Add newmatch to matched_pairs.
sourceIceFloeTracker.appendrows! — Methodappendrows!(df::MatchingProps, props::T, ratios, idx::Int64, dist::Float64) where {T<:DataFrameRow}
Append a row to df.props and df.ratios with the values of props and ratios respectively.
sourceIceFloeTracker.apply_cloudmask — Methodapply_cloudmask(reflectance_image, cloudmask)
Zero out pixels containing clouds where clouds and ice are not discernable. Arguments should be of the same size.
Arguments
reference_image: corrected reflectance false color image - bands [7,2,1] or grayscalecloudmask: binary cloudmask with clouds = 0, else = 1
sourceIceFloeTracker.apply_landmask — Methodapply_landmask(input_image, landmask_binary)
Zero out pixels in land and soft ice regions on truecolor image, return RGB image with zero for all three channels on land/soft ice.
Arguments
input_image: truecolor RGB imagelandmask_binary: binary landmask with 1=land, 0=water/ice
sourceIceFloeTracker.atan2 — Methodatan2(y,x)
Wrapper of Base.atan that returns the angle of vector (x,y) in the range [0, 2π).
sourceIceFloeTracker.binarize_landmask — Methodbinarize_landmask(landmask_image)
Convert a 3-channel RGB land mask image to a 1-channel binary matrix; land = 0, water/ice = 1.
Arguments
landmask_image: RGB land mask image from fetchdata
sourceIceFloeTracker.branch — Methodbranch(img::AbstractArray{Bool})
Find branch points in skeletonized image img according to Definition 3 of [1].
[1] Arcelli, Carlo, and Gabriella Sanniti di Baja. "Skeletons of planar patterns." Machine Intelligence and Pattern Recognition. Vol. 19. North-Holland, 1996. 99-143.
sourceIceFloeTracker.branch_candidates_func — Methodbranch_candidates_func(nhood)
Filter nhood as candidate for branch point.
To be passed to the make_lut function.
sourceIceFloeTracker.bridge — Methodbridge(bw)
Set 0-valued pixels to 1 if they have two nonzero neighbors that are not connected. Note the following exceptions:
0 0 0 0 0 0 1 0 1 becomes 1 1 1 0 0 0 0 0 0
1 0 1 1 1 1 0 0 0 becomes 0 0 0 0 0 0 0 0 0
The same applies to all their corresponding rotations.
Examples
julia> bw = [0 0 0; 0 0 0; 1 0 1]
+func2::Function=connected_background_count,)
Filter img with _operator_lut using lut1 and lut2.
sourceIceFloeTracker._operator_lut — Method_operator_lut(I, img, nhood, lut1, lut2)
Look up the neighborhood nhood in lookup tables lut1 and lut2.
Handles cases when the center of nhood is on the edge of img using data in I.
sourceIceFloeTracker.absdiffmeanratio — Methodabsdiffmeanratio(x, y)
Calculate the absolute difference between x and y divided by the mean of x and y.
sourceIceFloeTracker.add_padding — Methodadd_padding(img, style)
Extrapolate the image img according to the style specifications type. Returns the extrapolated image.
Arguments
img: Image to be padded.style: A supported type (such as Pad or Fill) representing the extrapolation style. See the relevant documentation for details.
See also remove_padding
sourceIceFloeTracker.add_passtimes! — Methodadd_passtimes!(props, passtimes)
Add a column passtime to each DataFrame in props containing the time of the image in which the floes were captured.
Arguments
props: array of DataFrames containing floe properties.passtimes: array of DateTime objects containing the time of the image in which the floes were captured.
sourceIceFloeTracker.addfloemasks! — Methodaddfloearrays(props::DataFrame, floeimg::BitMatrix)
Add a column to props called floearray containing the cropped floe masks from floeimg.
sourceIceFloeTracker.addmatch! — Methodaddmatch!(matched_pairs, newmatch)
Add newmatch to matched_pairs.
sourceIceFloeTracker.appendrows! — Methodappendrows!(df::MatchingProps, props::T, ratios, idx::Int64, dist::Float64) where {T<:DataFrameRow}
Append a row to df.props and df.ratios with the values of props and ratios respectively.
sourceIceFloeTracker.apply_cloudmask — Methodapply_cloudmask(reflectance_image, cloudmask)
Zero out pixels containing clouds where clouds and ice are not discernable. Arguments should be of the same size.
Arguments
reference_image: corrected reflectance false color image - bands [7,2,1] or grayscalecloudmask: binary cloudmask with clouds = 0, else = 1
sourceIceFloeTracker.apply_landmask — Methodapply_landmask(input_image, landmask_binary)
Zero out pixels in land and soft ice regions on truecolor image, return RGB image with zero for all three channels on land/soft ice.
Arguments
input_image: truecolor RGB imagelandmask_binary: binary landmask with 1=land, 0=water/ice
sourceIceFloeTracker.atan2 — Methodatan2(y,x)
Wrapper of Base.atan that returns the angle of vector (x,y) in the range [0, 2π).
sourceIceFloeTracker.binarize_landmask — Methodbinarize_landmask(landmask_image)
Convert a 3-channel RGB land mask image to a 1-channel binary matrix; land = 0, water/ice = 1.
Arguments
landmask_image: RGB land mask image from fetchdata
sourceIceFloeTracker.branch — Methodbranch(img::AbstractArray{Bool})
Find branch points in skeletonized image img according to Definition 3 of [1].
[1] Arcelli, Carlo, and Gabriella Sanniti di Baja. "Skeletons of planar patterns." Machine Intelligence and Pattern Recognition. Vol. 19. North-Holland, 1996. 99-143.
sourceIceFloeTracker.branch_candidates_func — Methodbranch_candidates_func(nhood)
Filter nhood as candidate for branch point.
To be passed to the make_lut function.
sourceIceFloeTracker.bridge — Methodbridge(bw)
Set 0-valued pixels to 1 if they have two nonzero neighbors that are not connected. Note the following exceptions:
0 0 0 0 0 0 1 0 1 becomes 1 1 1 0 0 0 0 0 0
1 0 1 1 1 1 0 0 0 becomes 0 0 0 0 0 0 0 0 0
The same applies to all their corresponding rotations.
Examples
julia> bw = [0 0 0; 0 0 0; 1 0 1]
3×3 Matrix{Int64}:
0 0 0
0 0 0
@@ -40,7 +40,7 @@
3×3 BitMatrix:
1 1 1
1 1 1
- 1 1 1
sourceIceFloeTracker.bwareamaxfilt — Functionbwareamaxfilt(bwimg::AbstractArray{Bool}, conn)
Filter the smaller (by area) connected components in bwimg keeping the (assumed unique) largest.
Uses 8-pixel connectivity by default (conn=8). Use conn=4 for 4-pixel connectivity.
sourceIceFloeTracker.bwareamaxfilt! — Functionsource IceFloeTracker.bwdist — Methodbwdist(bwimg)
Distance transform for binary image bwdist.
sourceIceFloeTracker.bwperim — Methodbwperim(bwimg)
Locate the pixels at the boundary of objects in an binary image bwimg using 8-pixel connectivity.
Arguments
bwimg: Binary (black/white – 1/0) image
Examples
julia> A = zeros(Bool, 13, 16); A[2:6, 2:6] .= 1; A[4:8, 7:10] .= 1; A[10:12,13:15] .= 1; A[10:12,3:6] .= 1;
+ 1 1 1
sourceIceFloeTracker.bwareamaxfilt — Functionbwareamaxfilt(bwimg::AbstractArray{Bool}, conn)
Filter the smaller (by area) connected components in bwimg keeping the (assumed unique) largest.
Uses 8-pixel connectivity by default (conn=8). Use conn=4 for 4-pixel connectivity.
sourceIceFloeTracker.bwareamaxfilt! — Functionsource IceFloeTracker.bwdist — Methodbwdist(bwimg)
Distance transform for binary image bwdist.
sourceIceFloeTracker.bwperim — Methodbwperim(bwimg)
Locate the pixels at the boundary of objects in an binary image bwimg using 8-pixel connectivity.
Arguments
bwimg: Binary (black/white – 1/0) image
Examples
julia> A = zeros(Bool, 13, 16); A[2:6, 2:6] .= 1; A[4:8, 7:10] .= 1; A[10:12,13:15] .= 1; A[10:12,3:6] .= 1;
julia> A
13×16 Matrix{Bool}:
@@ -72,7 +72,7 @@
0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0
0 0 1 0 0 1 0 0 0 0 0 0 1 0 1 0
0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0
- 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
sourceIceFloeTracker.bwtraceboundary — Methodbwtraceboundary(image::Union{Matrix{Int64},Matrix{Float64},T};
+ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
sourceIceFloeTracker.bwtraceboundary — Methodbwtraceboundary(image::Union{Matrix{Int64},Matrix{Float64},T};
P0::Union{Tuple{Int,Int},CartesianIndex{2},Nothing}=nothing,
closed::Bool=true) where T<:AbstractMatrix{Bool}
Trace the boundary of objects in image
Background pixels are represented as zero. The algorithm traces the boundary counterclockwise and an initial point P0 can be specified. If more than one boundary is detected and an initial point is provided, the boundary that contains this point is returned as a vector of CartesianIndex types. Otherwise an array of vectors is returned with all the detected boundaries in image.
Arguments
image: image, preferably binary with one single object, whose objects' boundaries are to be traced.P0: initial point of a target boundary.closed: if true (default) makes the inital point of a boundary equal to the last point.
Example
julia> A = zeros(Int, 13, 16); A[2:6, 2:6] .= 1; A[4:8, 7:10] .= 1; A[10:12,13:15] .= 1; A[10:12,3:6] .= 1;
@@ -104,7 +104,7 @@
CartesianIndex(11, 15)
CartesianIndex(10, 15)
CartesianIndex(10, 14)
- CartesianIndex(10, 13)
sourceIceFloeTracker.callmatchcorr — Methodcallmatchcorr(conditions)
Condition to decide whether match_corr should be called.
sourceIceFloeTracker.check_fname — Functioncheck_fname(fname)
Checks fname does not exist in current directory; throws an assertion if this condition is false.
Arguments
fname: String object or Symbol to a reference to a String representing a path.
sourceIceFloeTracker.clockwise — MethodMake a clockwise turn from the dir direction
sourceIceFloeTracker.compute_ratios — Methodcompute_ratios((props_day1, r), (props_day2,s))
Compute the ratios of the floe properties between the rth floe in props_day1 and the sth floe in props_day2. Return a tuple of the ratios.
Arguments
props_day1: floe properties for day 1r: index of floe in props_day1props_day2: floe properties for day 2s: index of floe in props_day2
sourceIceFloeTracker.compute_ratios_conditions — Methodcompute_ratios_conditions((props_day1, r), (props_day2, s), delta_time, t)
Compute the conditions for a match between the rth floe in props_day1 and the sth floe in props_day2. Return a tuple of the conditions.
Arguments
props_day1: floe properties for day 1r: index of floe in props_day1props_day2: floe properties for day 2s: index of floe in props_day2delta_time: time elapsed from image day 1 to image day 2t: tuple of thresholds for elapsed time and distance. See pair_floes for details.
sourceIceFloeTracker.connected_background_count — Methodconnected_background_count(nhood)
Second lut generator for neighbor transform with diamond strel (4-neighborhood).
To be passed to the make_lut function.
sourceIceFloeTracker.corr — Methodcorr(f1,f2)
Return the correlation between the psi-s curves p1 and p2.
sourceIceFloeTracker.corr — Methodcorr(f1,f2)
Return the normalized cross-correlation between the psi-s curves p1 and p2.
sourceIceFloeTracker.counterclockwise — MethodMake a counterclockwise turn from the dir direction
sourceIceFloeTracker.create_cloudmask — Methodcreate_cloudmask(reflectance_image; prelim_threshold, band_7_threshold, band_2_threshold, ratio_lower, ratio_upper)
Convert a 3-channel false color reflectance image to a 1-channel binary matrix; clouds = 0, else = 1. Default thresholds are defined in the published Ice Floe Tracker article: Remote Sensing of the Environment 234 (2019) 111406.
Arguments
reflectance_image: corrected reflectance false color image - bands [7,2,1]prelim_threshold: threshold value used to identify clouds in band 7, N0f8(RGB intensity/255)band_7_threshold: threshold value used to identify cloud-ice in band 7, N0f8(RGB intensity/255)band_2_threshold: threshold value used to identify cloud-ice in band 2, N0f8(RGB intensity/255)ratio_lower: threshold value used to set lower ratio of cloud-ice in bands 7 and 2ratio_upper: threshold value used to set upper ratio of cloud-ice in bands 7 and 2
sourceIceFloeTracker.create_landmask — Methodcreate_landmask(landmask_image, struct_elem, fill_value_lower, fill_value_upper)
Convert a 3-channel RGB land mask image to a 1-channel binary matrix, including a buffer to extend the land over any soft ice regions; land = 0, water/ice = 1. Returns a named tuple with the dilated and non-dilated landmask.
Arguments
landmask_image: RGB land mask image from fetchdatastruct_elem: structuring element for dilation (optional)fill_value_lower: fill holes having at least these many pixels (optional)fill_value_upper: fill holes having at most these many pixels (optional)
sourceIceFloeTracker.cropfloe — Methodcropfloe(floesimg, props, i)
Crops the floe delimited by the bounding box data in props at index i from the floe image floesimg.
sourceIceFloeTracker.crosscorr — Methodr, lags = crosscorr(u::Vector{T},
+ CartesianIndex(10, 13)
sourceIceFloeTracker.callmatchcorr — Methodcallmatchcorr(conditions)
Condition to decide whether match_corr should be called.
sourceIceFloeTracker.check_fname — Functioncheck_fname(fname)
Checks fname does not exist in current directory; throws an assertion if this condition is false.
Arguments
fname: String object or Symbol to a reference to a String representing a path.
sourceIceFloeTracker.clockwise — MethodMake a clockwise turn from the dir direction
sourceIceFloeTracker.compute_ratios — Methodcompute_ratios((props_day1, r), (props_day2,s))
Compute the ratios of the floe properties between the rth floe in props_day1 and the sth floe in props_day2. Return a tuple of the ratios.
Arguments
props_day1: floe properties for day 1r: index of floe in props_day1props_day2: floe properties for day 2s: index of floe in props_day2
sourceIceFloeTracker.compute_ratios_conditions — Methodcompute_ratios_conditions((props_day1, r), (props_day2, s), delta_time, t)
Compute the conditions for a match between the rth floe in props_day1 and the sth floe in props_day2. Return a tuple of the conditions.
Arguments
props_day1: floe properties for day 1r: index of floe in props_day1props_day2: floe properties for day 2s: index of floe in props_day2delta_time: time elapsed from image day 1 to image day 2t: tuple of thresholds for elapsed time and distance. See pair_floes for details.
sourceIceFloeTracker.connected_background_count — Methodconnected_background_count(nhood)
Second lut generator for neighbor transform with diamond strel (4-neighborhood).
To be passed to the make_lut function.
sourceIceFloeTracker.corr — Methodcorr(f1,f2)
Return the correlation between the psi-s curves p1 and p2.
sourceIceFloeTracker.corr — Methodcorr(f1,f2)
Return the normalized cross-correlation between the psi-s curves p1 and p2.
sourceIceFloeTracker.counterclockwise — MethodMake a counterclockwise turn from the dir direction
sourceIceFloeTracker.create_cloudmask — Methodcreate_cloudmask(reflectance_image; prelim_threshold, band_7_threshold, band_2_threshold, ratio_lower, ratio_upper)
Convert a 3-channel false color reflectance image to a 1-channel binary matrix; clouds = 0, else = 1. Default thresholds are defined in the published Ice Floe Tracker article: Remote Sensing of the Environment 234 (2019) 111406.
Arguments
reflectance_image: corrected reflectance false color image - bands [7,2,1]prelim_threshold: threshold value used to identify clouds in band 7, N0f8(RGB intensity/255)band_7_threshold: threshold value used to identify cloud-ice in band 7, N0f8(RGB intensity/255)band_2_threshold: threshold value used to identify cloud-ice in band 2, N0f8(RGB intensity/255)ratio_lower: threshold value used to set lower ratio of cloud-ice in bands 7 and 2ratio_upper: threshold value used to set upper ratio of cloud-ice in bands 7 and 2
sourceIceFloeTracker.create_landmask — Methodcreate_landmask(landmask_image, struct_elem, fill_value_lower, fill_value_upper)
Convert a 3-channel RGB land mask image to a 1-channel binary matrix, including a buffer to extend the land over any soft ice regions; land = 0, water/ice = 1. Returns a named tuple with the dilated and non-dilated landmask.
Arguments
landmask_image: RGB land mask image from fetchdatastruct_elem: structuring element for dilation (optional)fill_value_lower: fill holes having at least these many pixels (optional)fill_value_upper: fill holes having at most these many pixels (optional)
sourceIceFloeTracker.cropfloe — Methodcropfloe(floesimg, props, i)
Crops the floe delimited by the bounding box data in props at index i from the floe image floesimg.
sourceIceFloeTracker.crosscorr — Methodr, lags = crosscorr(u::Vector{T},
v::Vector{T};
normalize::Bool=false,
padmode::Symbol=:longest)
Wrapper of DSP.xcorr with normalization (see https://docs.juliadsp.org/stable/convolutions/#DSP.xcorr)
Returns the pair (r, lags) with the cross correlation scores r and corresponding lags according to padmode.
Arguments
u,v: real vectors which could have unequal length.normalize: return normalized correlation scores (false by default).padmode: either :longest (default) or :none to control padding of shorter vector with zeros.
Examples
The example below builds two vectors, one a shifted version of the other, and computes various cross correlation scores.
julia> n = 1:5;
@@ -167,7 +167,7 @@
0.477211 4.0
0.229061 5.0
0.105402 6.0
-0.0411191 7.0
sourceIceFloeTracker.deleteallbut! — Methoddeleteallbut!(matched_pairs, idxs, keeper)
Delete all rows in matched_pairs except for the row with index keeper in idxs.
sourceIceFloeTracker.detect_move! — MethodWorkhorse function: Get all pixel coords for detected border.
sourceIceFloeTracker.discriminate_ice_water — Methoddiscriminate_ice_water(
+0.0411191 7.0
sourceIceFloeTracker.deleteallbut! — Methoddeleteallbut!(matched_pairs, idxs, keeper)
Delete all rows in matched_pairs except for the row with index keeper in idxs.
sourceIceFloeTracker.detect_move! — MethodWorkhorse function: Get all pixel coords for detected border.
sourceIceFloeTracker.discriminate_ice_water — Methoddiscriminate_ice_water(
reflectance_image::Matrix{RGB{Float64}},
normalized_image::Matrix{Gray{Float64}},
landmask_bitmatrix::T,
@@ -182,8 +182,8 @@
st_dev_thresh_upper::Float64=Float64(98.9 / 255),
clouds_ratio_threshold::Float64=0.02,
differ_threshold::Float64=0.6,
-nbins::Real=155
)
Generates an image with ice floes apparent after filtering and combining previously processed versions of reflectance and truecolor images from the same region of interest. Returns an image ready for segmentation to isolate floes.
Note: This function mutates the landmask object to avoid unnecessary memory allocation. If you need the original landmask, make a copy before passing it to this function. Example: discriminate_ice_water(reflectance_image, normalized_image, copy(landmask_bitmatrix), cloudmask_bitmatrix)
Arguments
reflectance_image: input image in false color reflectancenormalized_image: normalized version of true color imagelandmask_bitmatrix: landmask for region of interestcloudmask_bitmatrix: cloudmask for region of interestfloes_threshold: heuristic applied to original reflectance imagemask_clouds_lower: lower heuristic applied to mask out cloudsmask_clouds_upper: upper heuristic applied to mask out cloudskurt_thresh_lower: lower heuristic used to set pixel value threshold based on kurtosis in histogramkurt_thresh_upper: upper heuristic used to set pixel value threshold based on kurtosis in histogramskew_thresh: heuristic used to set pixel value threshold based on skewness in histogramst_dev_thresh_lower: lower heuristic used to set pixel value threshold based on standard deviation in histogramst_dev_thresh_upper: upper heuristic used to set pixel value threshold based on standard deviation in histogramclouds2_threshold: heuristic used to set pixel value threshold based on ratio of cloudsdiffer_threshold: heuristic used to calculate proportional intensity in histogramnbins: number of bins during histogram build
sourceIceFloeTracker.dist — Methoddist(p1, p2)
Return the distance between the points p1 and p2.
sourceIceFloeTracker.filt_except_label! — Methodfilt_except_label!(labeled_arr::Array{Int64, 2}, label::Int64)
In-place version of filt_except_label.
See also filt_except_label
sourceIceFloeTracker.filt_except_label — Methodfilt_except_label(labeled_arr::Array{Int64, 2}, label::Int64)
Make 0 all values in labeled_arr that are not equal to label.
See also filt_except_label!
sourceIceFloeTracker.find_ice_labels — Methodfind_ice_labels(reflectance_image, landmask; band_7_threshold, band_2_threshold, band_1_threshold, band_7_relaxed_threshold, band_1_relaxed_threshold, possible_ice_threshold)
Locate the pixels of obvious ice from false color reflectance image. Returns a binary mask with ice floes contrasted from background. Default thresholds are defined in the published Ice Floe Tracker article: Remote Sensing of the Environment 234 (2019) 111406.
Arguments
reflectance_image: corrected reflectance false color image - bands [7,2,1]landmask: bitmatrix landmask for region of interestband_7_threshold: threshold value used to identify ice in band 7, N0f8(RGB intensity/255)band_2_threshold: threshold value used to identify ice in band 2, N0f8(RGB intensity/255)band_1_threshold: threshold value used to identify ice in band 2, N0f8(RGB intensity/255)band_7_relaxed_threshold: threshold value used to identify ice in band 7 if not found on first pass, N0f8(RGB intensity/255)band_1_relaxed_threshold: threshold value used to identify ice in band 1 if not found on first pass, N0f8(RGB intensity/255)
sourceIceFloeTracker.find_reflectance_peaks — Methodfind_reflectance_peaks(reflectance_channel, possible_ice_threshold;)
Find histogram peaks in single channels of a reflectance image and return the second greatest peak. If needed, edges can be returned as the first object from build_histogram. Similarly, peak values can be returned as the second object from findmaxima.
Arguments
reflectance_channel: either band 2 or band 1 of false-color reflectance imagepossible_ice_threshold: threshold value used to identify ice if not found on first or second pass
sourceIceFloeTracker.fixzeroindexing! — Methodfixzeroindexing!(props::DataFrame, props_to_fix::Vector{T}) where T<:Union{Symbol,String}
Fix the zero-indexing of the props_to_fix columns in props by adding 1 to each element.
sourceIceFloeTracker.fname_ext_splice — Methodfname_ext_splice(fname, ext)
Join "fname" and "ext" with '.'.
sourceIceFloeTracker.fname_ext_split — Methodfname_ext_split(fname)
Split "fname.ext" into "fname" and "ext".
sourceIceFloeTracker.from_to — MethodGet index in Moore neigborhood representing the direction from the from pixel coords to the to pixel coords (see definition of dir_delta below).
Clockwise Moore neighborhood.
dir_delta = [CartesianIndex(-1, 0), CartesianIndex(-1, 1), CartesianIndex(0, 1), CartesianIndex(1, 1), CartesianIndex(1, 0), CartesianIndex(1, -1), CartesianIndex(0, -1), CartesianIndex(-1,-1)]
sourceIceFloeTracker.get_areas — Methodget_areas(labeled_arr::Array{T, 2})::Dict{T, Int} where T
Get the "areas" (count of pixels of a given label) of the connected components in labeled_arr.
Return a dictionary with the frequency distribution: label => countoflabel.
sourceIceFloeTracker.get_max_label — Methodget_max_label(d::Dict{Int, Int})
Get the label k in dictionary d for which d[k] is maximal.
sourceIceFloeTracker.getbboxcolumns — Methodgetbboxcolumns(props::DataFrame)
Return the column names of the bounding box columns in props.
sourceIceFloeTracker.getbestmatchdata — Methodgetbestmatchdata(idx, r, props_day1, matching_floes)
Collect the data for the best match between the rth floe in props_day1 and the idxth floe in matching_floes. Return a tuple of the floe properties for day 1 and day 2 and the ratios.
sourceIceFloeTracker.getcentroid — Methodgetcentroid(props_day::DataFrame, r)
Get the coordinates of the rth floe in props_day.
sourceIceFloeTracker.getcentroidcolumns — Methodgetcentroidcolumns(props::DataFrame)
Returns the column names of the centroid columns in props.
sourceIceFloeTracker.getcollisions — Methodgetcollisions(matchedpairs)
Get nonunique rows in matchedpairs.
sourceIceFloeTracker.getcollisionslocs — Methodgetcollisionslocs(df)
Return a vector of tuples of the row and the index of the row in df that has a collision with another row in df.
sourceIceFloeTracker.getfloemasks — Methodgetfloemasks(props::DataFrame, floeimg::BitMatrix)
Return a vector of cropped floe masks from floeimg using the bounding box data in props.
sourceIceFloeTracker.getidxmostminimumeverything — Methodgetidxmostminimumeverything(ratiosdf)
Return the index of the row in ratiosdf with the most minima across its columns. If there are multiple columns with the same minimum value, return the index of the first column with the minimum value. If ratiosdf is empty, return NaN.
sourceIceFloeTracker.getidxofrow — Methodgetidxofrow(rw0, df)
Return the indices of the rows in df that are equal to rw0.
sourceIceFloeTracker.getpropsday1day2 — Methodgetpropsday1day2(properties, dayidx::Int64)
Return the floe properties for day dayidx and day dayidx+1.
sourceIceFloeTracker.getsizecomparability — Methodgetsizecomparability(s1, s2)
Check if the size of two floes s1 and s2 are comparable. The size is defined as the product of the floe dimensions.
Arguments
s1: size of floe 1s2: size of floe 2
sourceIceFloeTracker.grad — Methoddx, dy = grad(A::Matrix{<:Number})
Make gradient vector field for the set of points with coordinates in the rows of the matrix A with x-coordinates down column 1 and y-coordinates down column 2. Return a tuple with dx and dy in that order.
sourceIceFloeTracker.grad — Methoddx, dy = grad(x::Vector{<:Number}, y::Vector{<:Number})
Make gradient vector field for the set of points with coordinates in vectors x and y. Return a tuple with dx and dy in that order.
sourceIceFloeTracker.hbreak! — Methodhbreak!(img::AbstractArray{Bool})
Inplace version of hbreak. See also hbreak.
sourceIceFloeTracker.hbreak — Methodhbreak(img::AbstractArray{Bool})
Remove H-connected pixels in the binary image img. See also hbreak! for an inplace version of this function.
Examples
```jldoctest; setup = :(using IceFloeTracker)
julia> h1 = trues(3,3); h1[[1 3], 2] .= false; h1 3×3 BitMatrix: 1 0 1 1 1 1 1 0 1
julia> h2 = trues(3,3); h2[2, [1 3]] .= false; h2 3×3 BitMatrix: 1 1 1 0 1 0 1 1 1
julia> hbreak!(h1); h1 # modify h1 inplace 3×3 BitMatrix: 1 0 1 1 0 1 1 0 1
julia> hbreak(h2) 3×3 BitMatrix: 1 1 1 0 0 0 1 1 1
sourceIceFloeTracker.imextendedmin — Methodimextendedmin(img)
Mimics MATLAB's imextendedmin function that computes the extended-minima transform, which is the regional minima of the H-minima transform. Regional minima are connected components of pixels with a constant intensity value. This function returns a transformed bitmatrix.
Arguments
img: image objecth: suppress minima below this depth thresholdconn: neighborhood connectivity; in 2D 1 = 4-neighborhood and 2 = 8-neighborhood
sourceIceFloeTracker.imsharpen — Functionimsharpen(truecolor_image, lambda, kappa, niters, nbins, rblocks, cblocks, clip, smoothing_param, intensity)
Sharpen truecolor_image.
Arguments
truecolor_image: input image in truecolorlandmask_no_dilate: landmask for region of interestlambda: speed of diffusion (0–0.25)kappa: conduction coefficient for diffusion (25–100)niters: number of iterations of diffusionnbins: number of bins during histogram equalizationrblocks: number of row blocks to divide input image during equalizationcblocks: number of column blocks to divide input image during equalizationclip: Thresholds for clipping histogram bins (0–1); values closer to one minimize contrast enhancement, values closer to zero maximize contrast enhancement smoothing_param: pixel radius for gaussian blurring (1–10)intensity: amount of sharpening to perform
sourceIceFloeTracker.imsharpen_gray — Methodimsharpen_gray(imgsharpened, landmask)
Apply landmask and return Gray type image in colorview for normalization.
sourceIceFloeTracker.isfloegoodmatch — Methodisfloegoodmatch(conditions, mct, area_under, corr)
Return true if the floes are a good match as per the set thresholds. Return false otherwise.
Arguments
conditions: tuple of booleans for evaluating the conditionsmct: tuple of thresholds for the match correlation testarea_under and corr: values returned by match_corr
sourceIceFloeTracker.isincountourlist — MethodCheck P is in countour list if so return the index of the contour that contains P, otherwise return false.
sourceIceFloeTracker.ismember — Methodismember(df1,df2)
Return a boolean array indicating whether each row in df1 is a member of df2.
sourceIceFloeTracker.kmeans_segmentation — Methodkmeans_segmentation(gray_image, ice_labels;)
Apply k-means segmentation to a gray image to isolate a cluster group representing sea ice. Returns a binary image with ice segmented from background.
Arguments
gray_image: output image from ice-water-discrimination.jl or gray ice floe leads image in segmentation_f.jlice_labels: vector if pixel coordinates output from find_ice_labels.jl
sourceIceFloeTracker.loadimg — Methodloadimg(; dir::String, fname::String)
Load an image from dir with filename fname into a matrix of Float64 values. Returns the loaded image.
sourceIceFloeTracker.make_filename — Methodmake_filename()
Makes default filename with timestamp.
sourceIceFloeTracker.make_hbreak_dict — Methodmake_hbreak_dict()
Build dict with the two versions of an H-connected 3x3 neighboorhood.
h1 = [1 0 1 1 1 1 1 0 1]
h2 = [1 1 1 0 1 0 1 1 1]
sourceIceFloeTracker.make_lut — Methodmake_lut(lutfunc::Function)
Generate lookup table (lut) for 3x3 neighborhoods according to lutfunc.
sourceIceFloeTracker.make_psi_s — Methodmake_psi_s(XY::Matrix{<:Number};rangeout::Bool=true,
-unwrap::Bool=true)
Alternate method of make_psi_s accepting input vectors x and y as a 2-column matrix [x y] in order to facillitate workflow (output from resample_boundary).
sourceIceFloeTracker.make_psi_s — Methodmake_psi_s(x::Vector{<:Number},
+nbins::Real=155
)
Generates an image with ice floes apparent after filtering and combining previously processed versions of reflectance and truecolor images from the same region of interest. Returns an image ready for segmentation to isolate floes.
Note: This function mutates the landmask object to avoid unnecessary memory allocation. If you need the original landmask, make a copy before passing it to this function. Example: discriminate_ice_water(reflectance_image, normalized_image, copy(landmask_bitmatrix), cloudmask_bitmatrix)
Arguments
reflectance_image: input image in false color reflectancenormalized_image: normalized version of true color imagelandmask_bitmatrix: landmask for region of interestcloudmask_bitmatrix: cloudmask for region of interestfloes_threshold: heuristic applied to original reflectance imagemask_clouds_lower: lower heuristic applied to mask out cloudsmask_clouds_upper: upper heuristic applied to mask out cloudskurt_thresh_lower: lower heuristic used to set pixel value threshold based on kurtosis in histogramkurt_thresh_upper: upper heuristic used to set pixel value threshold based on kurtosis in histogramskew_thresh: heuristic used to set pixel value threshold based on skewness in histogramst_dev_thresh_lower: lower heuristic used to set pixel value threshold based on standard deviation in histogramst_dev_thresh_upper: upper heuristic used to set pixel value threshold based on standard deviation in histogramclouds2_threshold: heuristic used to set pixel value threshold based on ratio of cloudsdiffer_threshold: heuristic used to calculate proportional intensity in histogramnbins: number of bins during histogram build
sourceIceFloeTracker.dist — Methoddist(p1, p2)
Return the distance between the points p1 and p2.
sourceIceFloeTracker.filt_except_label! — Methodfilt_except_label!(labeled_arr::Array{Int64, 2}, label::Int64)
In-place version of filt_except_label.
See also filt_except_label
sourceIceFloeTracker.filt_except_label — Methodfilt_except_label(labeled_arr::Array{Int64, 2}, label::Int64)
Make 0 all values in labeled_arr that are not equal to label.
See also filt_except_label!
sourceIceFloeTracker.find_ice_labels — Methodfind_ice_labels(reflectance_image, landmask; band_7_threshold, band_2_threshold, band_1_threshold, band_7_relaxed_threshold, band_1_relaxed_threshold, possible_ice_threshold)
Locate the pixels of obvious ice from false color reflectance image. Returns a binary mask with ice floes contrasted from background. Default thresholds are defined in the published Ice Floe Tracker article: Remote Sensing of the Environment 234 (2019) 111406.
Arguments
reflectance_image: corrected reflectance false color image - bands [7,2,1]landmask: bitmatrix landmask for region of interestband_7_threshold: threshold value used to identify ice in band 7, N0f8(RGB intensity/255)band_2_threshold: threshold value used to identify ice in band 2, N0f8(RGB intensity/255)band_1_threshold: threshold value used to identify ice in band 2, N0f8(RGB intensity/255)band_7_relaxed_threshold: threshold value used to identify ice in band 7 if not found on first pass, N0f8(RGB intensity/255)band_1_relaxed_threshold: threshold value used to identify ice in band 1 if not found on first pass, N0f8(RGB intensity/255)
sourceIceFloeTracker.find_reflectance_peaks — Methodfind_reflectance_peaks(reflectance_channel, possible_ice_threshold;)
Find histogram peaks in single channels of a reflectance image and return the second greatest peak. If needed, edges can be returned as the first object from build_histogram. Similarly, peak values can be returned as the second object from findmaxima.
Arguments
reflectance_channel: either band 2 or band 1 of false-color reflectance imagepossible_ice_threshold: threshold value used to identify ice if not found on first or second pass
sourceIceFloeTracker.fixzeroindexing! — Methodfixzeroindexing!(props::DataFrame, props_to_fix::Vector{T}) where T<:Union{Symbol,String}
Fix the zero-indexing of the props_to_fix columns in props by adding 1 to each element.
sourceIceFloeTracker.fname_ext_splice — Methodfname_ext_splice(fname, ext)
Join "fname" and "ext" with '.'.
sourceIceFloeTracker.fname_ext_split — Methodfname_ext_split(fname)
Split "fname.ext" into "fname" and "ext".
sourceIceFloeTracker.from_to — MethodGet index in Moore neigborhood representing the direction from the from pixel coords to the to pixel coords (see definition of dir_delta below).
Clockwise Moore neighborhood.
dir_delta = [CartesianIndex(-1, 0), CartesianIndex(-1, 1), CartesianIndex(0, 1), CartesianIndex(1, 1), CartesianIndex(1, 0), CartesianIndex(1, -1), CartesianIndex(0, -1), CartesianIndex(-1,-1)]
sourceIceFloeTracker.get_areas — Methodget_areas(labeled_arr::Array{T, 2})::Dict{T, Int} where T
Get the "areas" (count of pixels of a given label) of the connected components in labeled_arr.
Return a dictionary with the frequency distribution: label => countoflabel.
sourceIceFloeTracker.get_max_label — Methodget_max_label(d::Dict{Int, Int})
Get the label k in dictionary d for which d[k] is maximal.
sourceIceFloeTracker.getbboxcolumns — Methodgetbboxcolumns(props::DataFrame)
Return the column names of the bounding box columns in props.
sourceIceFloeTracker.getbestmatchdata — Methodgetbestmatchdata(idx, r, props_day1, matching_floes)
Collect the data for the best match between the rth floe in props_day1 and the idxth floe in matching_floes. Return a tuple of the floe properties for day 1 and day 2 and the ratios.
sourceIceFloeTracker.getcentroid — Methodgetcentroid(props_day::DataFrame, r)
Get the coordinates of the rth floe in props_day.
sourceIceFloeTracker.getcentroidcolumns — Methodgetcentroidcolumns(props::DataFrame)
Returns the column names of the centroid columns in props.
sourceIceFloeTracker.getcollisions — Methodgetcollisions(matchedpairs)
Get nonunique rows in matchedpairs.
sourceIceFloeTracker.getcollisionslocs — Methodgetcollisionslocs(df)
Return a vector of tuples of the row and the index of the row in df that has a collision with another row in df.
sourceIceFloeTracker.getfloemasks — Methodgetfloemasks(props::DataFrame, floeimg::BitMatrix)
Return a vector of cropped floe masks from floeimg using the bounding box data in props.
sourceIceFloeTracker.getidxmostminimumeverything — Methodgetidxmostminimumeverything(ratiosdf)
Return the index of the row in ratiosdf with the most minima across its columns. If there are multiple columns with the same minimum value, return the index of the first column with the minimum value. If ratiosdf is empty, return NaN.
sourceIceFloeTracker.getidxofrow — Methodgetidxofrow(rw0, df)
Return the indices of the rows in df that are equal to rw0.
sourceIceFloeTracker.getpropsday1day2 — Methodgetpropsday1day2(properties, dayidx::Int64)
Return the floe properties for day dayidx and day dayidx+1.
sourceIceFloeTracker.getsizecomparability — Methodgetsizecomparability(s1, s2)
Check if the size of two floes s1 and s2 are comparable. The size is defined as the product of the floe dimensions.
Arguments
s1: size of floe 1s2: size of floe 2
sourceIceFloeTracker.grad — Methoddx, dy = grad(A::Matrix{<:Number})
Make gradient vector field for the set of points with coordinates in the rows of the matrix A with x-coordinates down column 1 and y-coordinates down column 2. Return a tuple with dx and dy in that order.
sourceIceFloeTracker.grad — Methoddx, dy = grad(x::Vector{<:Number}, y::Vector{<:Number})
Make gradient vector field for the set of points with coordinates in vectors x and y. Return a tuple with dx and dy in that order.
sourceIceFloeTracker.hbreak! — Methodhbreak!(img::AbstractArray{Bool})
Inplace version of hbreak. See also hbreak.
sourceIceFloeTracker.hbreak — Methodhbreak(img::AbstractArray{Bool})
Remove H-connected pixels in the binary image img. See also hbreak! for an inplace version of this function.
Examples
```jldoctest; setup = :(using IceFloeTracker)
julia> h1 = trues(3,3); h1[[1 3], 2] .= false; h1 3×3 BitMatrix: 1 0 1 1 1 1 1 0 1
julia> h2 = trues(3,3); h2[2, [1 3]] .= false; h2 3×3 BitMatrix: 1 1 1 0 1 0 1 1 1
julia> hbreak!(h1); h1 # modify h1 inplace 3×3 BitMatrix: 1 0 1 1 0 1 1 0 1
julia> hbreak(h2) 3×3 BitMatrix: 1 1 1 0 0 0 1 1 1
sourceIceFloeTracker.imextendedmin — Methodimextendedmin(img)
Mimics MATLAB's imextendedmin function that computes the extended-minima transform, which is the regional minima of the H-minima transform. Regional minima are connected components of pixels with a constant intensity value. This function returns a transformed bitmatrix.
Arguments
img: image objecth: suppress minima below this depth thresholdconn: neighborhood connectivity; in 2D 1 = 4-neighborhood and 2 = 8-neighborhood
sourceIceFloeTracker.imsharpen — Functionimsharpen(truecolor_image, lambda, kappa, niters, nbins, rblocks, cblocks, clip, smoothing_param, intensity)
Sharpen truecolor_image.
Arguments
truecolor_image: input image in truecolorlandmask_no_dilate: landmask for region of interestlambda: speed of diffusion (0–0.25)kappa: conduction coefficient for diffusion (25–100)niters: number of iterations of diffusionnbins: number of bins during histogram equalizationrblocks: number of row blocks to divide input image during equalizationcblocks: number of column blocks to divide input image during equalizationclip: Thresholds for clipping histogram bins (0–1); values closer to one minimize contrast enhancement, values closer to zero maximize contrast enhancement smoothing_param: pixel radius for gaussian blurring (1–10)intensity: amount of sharpening to perform
sourceIceFloeTracker.imsharpen_gray — Methodimsharpen_gray(imgsharpened, landmask)
Apply landmask and return Gray type image in colorview for normalization.
sourceIceFloeTracker.isfloegoodmatch — Methodisfloegoodmatch(conditions, mct, area_under, corr)
Return true if the floes are a good match as per the set thresholds. Return false otherwise.
Arguments
conditions: tuple of booleans for evaluating the conditionsmct: tuple of thresholds for the match correlation testarea_under and corr: values returned by match_corr
sourceIceFloeTracker.isincountourlist — MethodCheck P is in countour list if so return the index of the contour that contains P, otherwise return false.
sourceIceFloeTracker.ismember — Methodismember(df1,df2)
Return a boolean array indicating whether each row in df1 is a member of df2.
sourceIceFloeTracker.kmeans_segmentation — Methodkmeans_segmentation(gray_image, ice_labels;)
Apply k-means segmentation to a gray image to isolate a cluster group representing sea ice. Returns a binary image with ice segmented from background.
Arguments
gray_image: output image from ice-water-discrimination.jl or gray ice floe leads image in segmentation_f.jlice_labels: vector if pixel coordinates output from find_ice_labels.jl
sourceIceFloeTracker.loadimg — Methodloadimg(; dir::String, fname::String)
Load an image from dir with filename fname into a matrix of Float64 values. Returns the loaded image.
sourceIceFloeTracker.make_filename — Methodmake_filename()
Makes default filename with timestamp.
sourceIceFloeTracker.make_hbreak_dict — Methodmake_hbreak_dict()
Build dict with the two versions of an H-connected 3x3 neighboorhood.
h1 = [1 0 1 1 1 1 1 0 1]
h2 = [1 1 1 0 1 0 1 1 1]
sourceIceFloeTracker.make_lut — Methodmake_lut(lutfunc::Function)
Generate lookup table (lut) for 3x3 neighborhoods according to lutfunc.
sourceIceFloeTracker.make_psi_s — Methodmake_psi_s(XY::Matrix{<:Number};rangeout::Bool=true,
+unwrap::Bool=true)
Alternate method of make_psi_s accepting input vectors x and y as a 2-column matrix [x y] in order to facillitate workflow (output from resample_boundary).
sourceIceFloeTracker.make_psi_s — Methodmake_psi_s(x::Vector{<:Number},
y::Vector{<:Number};
rangeout::Bool=true,
unwrap::Bool=true)::Tuple{Vector{Float64}, Vector{Float64}}
Builds the ψ-s curve defined by vectors x and y.
Returns a tuple of vectors with the phases ψ in the first component and the traversed arclength in the second component.
Following the convention in [1], the wrapped ψ-s curve has values in [0, 2π) by default; use rangeout to control this behavior.
See also bwtraceboundary, resample_boundary
Arguments
x: Vector of x-coordinatesy: corresponding vector of y-coordinatesrangeout: true (default) for phase values in [0, 2π); false for phase values in (-π, π].unwrap: set to true to get "unwrapped" phases (default).
Reference
[1] McConnell, Ross, et al. "psi-s correlation and dynamic time warping: two methods for tracking ice floes in SAR images." IEEE Transactions on Geoscience and Remote sensing 29.6 (1991): 1004-1012.
Example
The example below builds a cardioid and obtains its ψ-s curve.
julia> t = range(0,2pi,201);
@@ -219,7 +219,7 @@
7.99877 9.35147
7.99926 9.39336
- julia> plot(s, psi) # inspect psi-s curve -- should be a straight line from (0, 0) to (8, 3π)
sourceIceFloeTracker.makecontourstartatP — MethodMake contour start at point P by permuting the elements in contour.
sourceIceFloeTracker.makeemptydffrom — Methodmakeemptydffrom(df::DataFrame)
Return an object with an empty dataframe with the same column names as df and an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios.
sourceIceFloeTracker.makeemptyratiosdf — Methodmakeemptyratiosdf()
Return an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios.
sourceIceFloeTracker.matchcorr — Methodmatchcorr(
+ julia> plot(s, psi) # inspect psi-s curve -- should be a straight line from (0, 0) to (8, 3π)
sourceIceFloeTracker.makecontourstartatP — MethodMake contour start at point P by permuting the elements in contour.
sourceIceFloeTracker.makeemptydffrom — Methodmakeemptydffrom(df::DataFrame)
Return an object with an empty dataframe with the same column names as df and an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios.
sourceIceFloeTracker.makeemptyratiosdf — Methodmakeemptyratiosdf()
Return an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios.
sourceIceFloeTracker.matchcorr — Methodmatchcorr(
f1::T,
f2::T,
Δt::F,
@@ -229,18 +229,18 @@
comp::F=0.25,
mm::F=0.22
)
-where {T<:AbstractArray{Bool,2},S<:Int64,F<:Float64}
Compute the mismatch mm and psi-s-correlation c for floes with masks f1 and f2.
The criteria for floes to be considered equivalent is as follows: - c greater than mm - _mm is less than mm
A pair of NaN is returned for cases for which one of their mask dimension is too small or their sizes are not comparable.
Arguments
f1: mask of floe 1f2: mask of floe 2Δt: time difference between floesmxrot: maximum rotation (in degrees) allowed between floesn (default: 10)psi: psi-s-correlation threshold (default: 0.95)sz: size threshold (default: 16)comp: size comparability threshold (default: 0.25)mm: mismatch threshold (default: 0.22)
sourceIceFloeTracker.mean — Methodmean(x,y)
Compute the mean of x and y.
sourceIceFloeTracker.mismatch — Functionmismatch(fixed::AbstractArray,
+where {T<:AbstractArray{Bool,2},S<:Int64,F<:Float64}
Compute the mismatch mm and psi-s-correlation c for floes with masks f1 and f2.
The criteria for floes to be considered equivalent is as follows: - c greater than mm - _mm is less than mm
A pair of NaN is returned for cases for which one of their mask dimension is too small or their sizes are not comparable.
Arguments
f1: mask of floe 1f2: mask of floe 2Δt: time difference between floesmxrot: maximum rotation (in degrees) allowed between floesn (default: 10)psi: psi-s-correlation threshold (default: 0.95)sz: size threshold (default: 16)comp: size comparability threshold (default: 0.25)mm: mismatch threshold (default: 0.22)
sourceIceFloeTracker.mean — Methodmean(x,y)
Compute the mean of x and y.
sourceIceFloeTracker.mismatch — Functionmismatch(fixed::AbstractArray,
moving::AbstractArray,
mxshift::Tuple{Int64, Int64}=(10,10),
mxrot::Float64=pi/4;
kwargs...
- )
Estimate a rigid transformation (translation + rotation) that minimizes the 'mismatch' of aligning moving with fixed using the QuadDIRECT algorithm.
Returns a pair with the mismatch score mm and the associated registration angle rot.
Arguments
fixed,moving: images to align via a rigid transformationmxshift: maximum allowed translation in units of array indices (default set to (10,10))mxrot: maximum allowed rotation in radians (default set to π/4)thresh: minimum sum-of-squared-intensity overlap between the images (default is 10% of the sum-of-squared-intensity of fixed)kwargs: other arguments such as tol, ftol, and fvalue (see QuadDIRECT.analyze for details)
```
sourceIceFloeTracker.modenan — Methodmodenan(x::AbstractVector{<:Float64})
Return the mode of x or NaN if x is empty.
sourceIceFloeTracker.move — Methodmove from current pixel to the next in given direction
sourceIceFloeTracker.norm — Methodnorm(v)
Get the euclidean norm of the vector v.
sourceIceFloeTracker.normalize_image — Methodnormalize_image(image_sharpened, image_sharpened_gray, landmask, struct_elem;)
Adjusts sharpened land-masked image to highlight ice floe features.
Does reconstruction and landmasking to image_sharpened.
Arguments
image_sharpened: sharpened image (output of imsharpen)image_sharpened_gray: grayscale, landmasked sharpened image (output of imsharpen_gray(image_sharpened))landmask: landmask for region of intereststruct_elem: structuring element for dilation
sourceIceFloeTracker.normalizeangle — Functionnormalizeangle(revised,t=180)
Normalize angle to be between -180 and 180 degrees.
sourceIceFloeTracker.padnhood — Methodpadnhood(img, I, nhood)
Pad the matrix img[nhood] with zeros according to the position of I within the edgesimg.
Returns img[nhood] if I is not an edge index.
sourceIceFloeTracker.pairfloes — Methodpairfloes(
+ )
Estimate a rigid transformation (translation + rotation) that minimizes the 'mismatch' of aligning moving with fixed using the QuadDIRECT algorithm.
Returns a pair with the mismatch score mm and the associated registration angle rot.
Arguments
fixed,moving: images to align via a rigid transformationmxshift: maximum allowed translation in units of array indices (default set to (10,10))mxrot: maximum allowed rotation in radians (default set to π/4)thresh: minimum sum-of-squared-intensity overlap between the images (default is 10% of the sum-of-squared-intensity of fixed)kwargs: other arguments such as tol, ftol, and fvalue (see QuadDIRECT.analyze for details)
```
sourceIceFloeTracker.modenan — Methodmodenan(x::AbstractVector{<:Float64})
Return the mode of x or NaN if x is empty.
sourceIceFloeTracker.move — Methodmove from current pixel to the next in given direction
sourceIceFloeTracker.norm — Methodnorm(v)
Get the euclidean norm of the vector v.
sourceIceFloeTracker.normalize_image — Methodnormalize_image(image_sharpened, image_sharpened_gray, landmask, struct_elem;)
Adjusts sharpened land-masked image to highlight ice floe features.
Does reconstruction and landmasking to image_sharpened.
Arguments
image_sharpened: sharpened image (output of imsharpen)image_sharpened_gray: grayscale, landmasked sharpened image (output of imsharpen_gray(image_sharpened))landmask: landmask for region of intereststruct_elem: structuring element for dilation
sourceIceFloeTracker.normalizeangle — Functionnormalizeangle(revised,t=180)
Normalize angle to be between -180 and 180 degrees.
sourceIceFloeTracker.padnhood — Methodpadnhood(img, I, nhood)
Pad the matrix img[nhood] with zeros according to the position of I within the edgesimg.
Returns img[nhood] if I is not an edge index.
sourceIceFloeTracker.pairfloes — Methodpairfloes(
segmented_imgs::Vector{BitMatrix},
props::Vector{DataFrame},
passtimes::Vector{DateTime},
dt::Vector{Float64},
condition_thresholds,
-mc_thresholds,
)
Pair floes in props[k] to floes in props[k+1] for k=1:length(props)-1.
The main steps of the algorithm are as follows:
- Crop floes from
segmented_imgs using bounding box data in props. Floes in the edges are removed. - For each floekr in
props[k], compare to floek+1s in props[k+1] by computing similarity ratios, set of conditions, and drift distance dist. If the conditions are met, compute the area mismatch mm and psi-s correlation c for this pair of floes. Pair these two floes if mm and c satisfy the thresholds in mc_thresholds. - If there are collisions (i.e. floe
s in props[k+1] is paired to more than one floe in props[k]), then the floe in props[k] with the best match is paired to floe s in props[k+1]. - Drop paired floes from
props[k] and props[k+1] and repeat steps 2 and 3 until there are no more floes to match in props[k]. - Repeat steps 2-4 for
k=2:length(props)-1.
Arguments
segmented_imgs: array of images with segmented floes.props: array of dataframes containing floe properties.passtimes: array of DateTime objects containing the time of the image in which the floes were captured.dt: array of time elapsed between images in segmented_imgs.condition_thresholds: 3-tuple of thresholds (each a named tuple) for deciding whether to match floe i from day k to floe j from day k+1.mc_thresholds: thresholds for area mismatch and psi-s shape correlation.
Returns a tuple (props, trackdata) where props is a long dataframe containing floe ID's, passtimes, the original set of physical properties, and their masks and trackdata is a dataframe containing the floe tracking data.
sourceIceFloeTracker.regionprops — Functionregionprops(label_img, ; properties, connectivity)
A wrapper of the regionprops function from the skimage python library.
See its full documentation at https://scikit-image.org/docs/stable/api/skimage.measure.html#skimage.measure.regionprops.
Arguments
label_img: Image with the labeled objects of interestintensity_img: (Optional) Used for generating extra_properties, integer/float array from which (presumably) label_img was generatedextra_properties: (Optional) not yet implemented. It will be set to nothing
See also regionprops_table
Examples
julia> Random.seed!(123);
+mc_thresholds,
)
Pair floes in props[k] to floes in props[k+1] for k=1:length(props)-1.
The main steps of the algorithm are as follows:
- Crop floes from
segmented_imgs using bounding box data in props. Floes in the edges are removed. - For each floekr in
props[k], compare to floek+1s in props[k+1] by computing similarity ratios, set of conditions, and drift distance dist. If the conditions are met, compute the area mismatch mm and psi-s correlation c for this pair of floes. Pair these two floes if mm and c satisfy the thresholds in mc_thresholds. - If there are collisions (i.e. floe
s in props[k+1] is paired to more than one floe in props[k]), then the floe in props[k] with the best match is paired to floe s in props[k+1]. - Drop paired floes from
props[k] and props[k+1] and repeat steps 2 and 3 until there are no more floes to match in props[k]. - Repeat steps 2-4 for
k=2:length(props)-1.
Arguments
segmented_imgs: array of images with segmented floes.props: array of dataframes containing floe properties.passtimes: array of DateTime objects containing the time of the image in which the floes were captured.dt: array of time elapsed between images in segmented_imgs.condition_thresholds: 3-tuple of thresholds (each a named tuple) for deciding whether to match floe i from day k to floe j from day k+1.mc_thresholds: thresholds for area mismatch and psi-s shape correlation.
Returns a tuple (props, trackdata) where props is a long dataframe containing floe ID's, passtimes, the original set of physical properties, and their masks and trackdata is a dataframe containing the floe tracking data.
sourceIceFloeTracker.regionprops — Functionregionprops(label_img, ; properties, connectivity)
A wrapper of the regionprops function from the skimage python library.
See its full documentation at https://scikit-image.org/docs/stable/api/skimage.measure.html#skimage.measure.regionprops.
Arguments
label_img: Image with the labeled objects of interestintensity_img: (Optional) Used for generating extra_properties, integer/float array from which (presumably) label_img was generatedextra_properties: (Optional) not yet implemented. It will be set to nothing
See also regionprops_table
Examples
julia> Random.seed!(123);
julia> bw_img = rand([0, 1], 5, 10)
5×10 Matrix{Int64}:
@@ -266,7 +266,7 @@
13 11.621320343559642
1 0.0
1 0.0
-7 4.621320343559642
sourceIceFloeTracker.regionprops_table — Functionregionprops_table(label_img, intensity_img; properties, connectivity, extra_properties)
A wrapper of the regionprops_table function from the skimage python library.
See its full documentation at https://scikit-image.org/docs/stable/api/skimage.measure.html#regionprops-table.
Arguments
label_img: Image with the labeled objects of interestintensity_img: (Optional) Used for generating extra_properties, integer/float array from which (presumably) label_img was generated properties: List (Vector or Tuple) of properties to be generated for each connected component in label_imgextra_properties: (Optional) not yet implemented. It will be set to nothing
Notes
- Zero indexing has been corrected for the
bbox and centroid properties bbox data (max_col and max_row) are inclusivecentroid data are rounded to the nearest integer
See also regionprops
Examples
julia> using IceFloeTracker, Random
+7 4.621320343559642
sourceIceFloeTracker.regionprops_table — Functionregionprops_table(label_img, intensity_img; properties, connectivity, extra_properties)
A wrapper of the regionprops_table function from the skimage python library.
See its full documentation at https://scikit-image.org/docs/stable/api/skimage.measure.html#regionprops-table.
Arguments
label_img: Image with the labeled objects of interestintensity_img: (Optional) Used for generating extra_properties, integer/float array from which (presumably) label_img was generated properties: List (Vector or Tuple) of properties to be generated for each connected component in label_imgextra_properties: (Optional) not yet implemented. It will be set to nothing
Notes
- Zero indexing has been corrected for the
bbox and centroid properties bbox data (max_col and max_row) are inclusivecentroid data are rounded to the nearest integer
See also regionprops
Examples
julia> using IceFloeTracker, Random
julia> Random.seed!(123);
@@ -299,16 +299,16 @@
1 │ 13 11.6213
2 │ 1 0.0
3 │ 1 0.0
- 4 │ 7 4.62132
sourceIceFloeTracker.remove_padding — Methodremove_padding(paddedimg, border_spec)
Removes padding from the boundary of padded image paddedimg according to the border specification border_spec type. Returns the cropped image.
Arguments
paddedimg: Pre-padded image.border_spec: Type representing the style of padding (such as Pad or Fill) with which paddedimg is assumend to be pre-padded. Example: Pad((1,2), (3,4)) specifies 1 row on the top, 2 columns on the left, 3 rows on the bottom, and 4 columns on the right boundary.
See also add_padding
sourceIceFloeTracker.renamecols! — Methodrenamecols!(props::DataFrame, oldnames::Vector{T}, newnames::Vector{T}) where T<:Union{Symbol,String}
Rename the oldnames columns in props to newnames.
sourceIceFloeTracker.resample_boundary — Functionresample_boundary(bd_points::Vector{<:CartesianIndex}, reduc_factor::Int64=2, bd::String="natural")
Get a uniform set of resampled boundary points from bd_points using cubic splines with specified boundary conditions
The resampled set of points is obtained using parametric interpolation of the points in bd_points. It is assumed that the separation between a pair of adjacent points is 1.
Arguments
bd_points: Sequetial set of boundary points for the object of interestreduc_factor: Factor by which to reduce the number of points in bd_points (2 by default)
-bd: Boundary condition, either 'natural' (default) or 'periodic'
See also bwtraceboundary
Example
```jldoctest; setup = :(using IceFloeTracker) julia> A = zeros(Int, 13, 16); A[2:6, 2:6] .= 1; A[4:8, 7:10] .= 1; A[10:12,13:15] .= 1; A[10:12,3:6] .= 1; julia> A 13×16 Matrix{Int64}: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
julia> boundary = bwtraceboundary(A);
julia> boundary[3] 9-element Vector{CartesianIndex}: CartesianIndex(10, 13) CartesianIndex(11, 13) CartesianIndex(12, 13) CartesianIndex(12, 14) CartesianIndex(12, 15) CartesianIndex(11, 15) CartesianIndex(10, 15) CartesianIndex(10, 14) CartesianIndex(10, 13)
julia> resample_boundary(boundary[3]) 4×2 Matrix{Float64}: 10.0 13.0 12.0357 13.5859 10.5859 15.0357 10.0 13.0
sourceIceFloeTracker.roundtoint! — Methodroundtoint!(props::DataFrame, colnames::Vector{T}) where T<:Union{Symbol,String}
Round the colnames columns in props to Int.
sourceIceFloeTracker.segmentation_A — Methodsegmentation_A(segmented_ice_cloudmasked; min_opening_area)
Apply k-means segmentation to a gray image to isolate a cluster group representing sea ice. Returns an image segmented and processed as well as an intermediate files needed for downstream functions.
Arguments
segmented_ice_cloudmask: bitmatrix with open water/clouds = 0, ice = 1, output from segmented_ice_cloudmasking()min_opening_area: minimum size of pixels to use during morphological openingfill_range: range of values dictating the size of holes to fill
sourceIceFloeTracker.segmentation_B — Functionsegmentation_B(sharpened_image, cloudmask, segmented_a_ice_mask, struct_elem; fill_range, isolation_threshold, alpha_level, adjusted_ice_threshold)
Performs image processing and morphological filtering with intermediate files from normalization.jl and segmentation_A to further isolate ice floes, returning a mask of potential ice.
Arguments
sharpened_image: non-cloudmasked but sharpened image, output from normalization.jlcloudmask: bitmatrix cloudmask for region of interestsegmented_a_ice_mask: binary cloudmasked ice mask from segmentation_a_direct.jlstruct_elem: structuring element for dilationfill_range: range of values dictating the size of holes to fillisolation_threshold: threshold used to isolated pixels from sharpened_image; between 0-1alpha_level: alpha threshold used to adjust contrastgamma_factor: amount of gamma adjustment adjusted_ice_threshold: threshold used to set ice equal to one after gamma adjustment
sourceIceFloeTracker.segmentation_F — Methodsegmentation_F(
+ 4 │ 7 4.62132
sourceIceFloeTracker.remove_padding — Methodremove_padding(paddedimg, border_spec)
Removes padding from the boundary of padded image paddedimg according to the border specification border_spec type. Returns the cropped image.
Arguments
paddedimg: Pre-padded image.border_spec: Type representing the style of padding (such as Pad or Fill) with which paddedimg is assumend to be pre-padded. Example: Pad((1,2), (3,4)) specifies 1 row on the top, 2 columns on the left, 3 rows on the bottom, and 4 columns on the right boundary.
See also add_padding
sourceIceFloeTracker.renamecols! — Methodrenamecols!(props::DataFrame, oldnames::Vector{T}, newnames::Vector{T}) where T<:Union{Symbol,String}
Rename the oldnames columns in props to newnames.
sourceIceFloeTracker.resample_boundary — Functionresample_boundary(bd_points::Vector{<:CartesianIndex}, reduc_factor::Int64=2, bd::String="natural")
Get a uniform set of resampled boundary points from bd_points using cubic splines with specified boundary conditions
The resampled set of points is obtained using parametric interpolation of the points in bd_points. It is assumed that the separation between a pair of adjacent points is 1.
Arguments
bd_points: Sequetial set of boundary points for the object of interestreduc_factor: Factor by which to reduce the number of points in bd_points (2 by default)
-bd: Boundary condition, either 'natural' (default) or 'periodic'
See also bwtraceboundary
Example
```jldoctest; setup = :(using IceFloeTracker) julia> A = zeros(Int, 13, 16); A[2:6, 2:6] .= 1; A[4:8, 7:10] .= 1; A[10:12,13:15] .= 1; A[10:12,3:6] .= 1; julia> A 13×16 Matrix{Int64}: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
julia> boundary = bwtraceboundary(A);
julia> boundary[3] 9-element Vector{CartesianIndex}: CartesianIndex(10, 13) CartesianIndex(11, 13) CartesianIndex(12, 13) CartesianIndex(12, 14) CartesianIndex(12, 15) CartesianIndex(11, 15) CartesianIndex(10, 15) CartesianIndex(10, 14) CartesianIndex(10, 13)
julia> resample_boundary(boundary[3]) 4×2 Matrix{Float64}: 10.0 13.0 12.0357 13.5859 10.5859 15.0357 10.0 13.0
sourceIceFloeTracker.roundtoint! — Methodroundtoint!(props::DataFrame, colnames::Vector{T}) where T<:Union{Symbol,String}
Round the colnames columns in props to Int.
sourceIceFloeTracker.segmentation_A — Methodsegmentation_A(segmented_ice_cloudmasked; min_opening_area)
Apply k-means segmentation to a gray image to isolate a cluster group representing sea ice. Returns an image segmented and processed as well as an intermediate files needed for downstream functions.
Arguments
segmented_ice_cloudmask: bitmatrix with open water/clouds = 0, ice = 1, output from segmented_ice_cloudmasking()min_opening_area: minimum size of pixels to use during morphological openingfill_range: range of values dictating the size of holes to fill
sourceIceFloeTracker.segmentation_B — Functionsegmentation_B(sharpened_image, cloudmask, segmented_a_ice_mask, struct_elem; fill_range, isolation_threshold, alpha_level, adjusted_ice_threshold)
Performs image processing and morphological filtering with intermediate files from normalization.jl and segmentation_A to further isolate ice floes, returning a mask of potential ice.
Arguments
sharpened_image: non-cloudmasked but sharpened image, output from normalization.jlcloudmask: bitmatrix cloudmask for region of interestsegmented_a_ice_mask: binary cloudmasked ice mask from segmentation_a_direct.jlstruct_elem: structuring element for dilationfill_range: range of values dictating the size of holes to fillisolation_threshold: threshold used to isolated pixels from sharpened_image; between 0-1alpha_level: alpha threshold used to adjust contrastgamma_factor: amount of gamma adjustment adjusted_ice_threshold: threshold used to set ice equal to one after gamma adjustment
sourceIceFloeTracker.segmentation_F — Methodsegmentation_F(
segmentation_B_not_ice_mask::Matrix{Gray{Float64}},
segmentation_B_ice_intersect::BitMatrix,
segmentation_B_watershed_intersect::BitMatrix,
ice_labels::Vector{Int64},
cloudmask::BitMatrix,
landmask::BitMatrix;
-min_area_opening::Int64=20
)
Cleans up past segmentation images with morphological operations, and applies the results of prior watershed segmentation, returning the final cleaned image for tracking with ice floes segmented and isolated.
Arguments
segmentation_B_not_ice_mask: gray image output from segmentation_b.jlsegmentation_B_ice_intersect: binary mask output from segmentation_b.jlsegmentation_B_watershed_intersect: ice pixels, output from segmentation_b.jl ice_labels: vector of pixel coordinates output from find_ice_labels.jlcloudmask.jl: bitmatrix cloudmask for region of interestlandmask.jl: bitmatrix landmask for region of interestmin_area_opening: threshold used for area opening; pixel groups greater than threshold are retained
sourceIceFloeTracker.segmented_ice_cloudmasking — Methodsegmented_ice_cloudmasking(gray_image, cloudmask, ice_labels;)
Apply cloudmask to a bitmatrix of segmented ice after kmeans clustering. Returns a bitmatrix with open water/clouds = 0, ice = 1).
Arguments
gray_image: output image from ice-water-discrimination.jl or gray ice floe leads image in segmentation_f.jlcloudmask: bitmatrix cloudmask for region of interestice_labels: vector if pixel coordinates output from find_ice_labels.jl
sourceIceFloeTracker.sort! — Methodsort!(tracked::Tracked)
Sort the floes in tracked by area in descending order.
sourceIceFloeTracker.sort_floes_by_area! — Methodsort_floes_by_area!(props)
Sort floes in props by area in descending order.
sourceIceFloeTracker.timestamp — Methodtimestamp(fname)
Attach timestamp to fname.
sourceIceFloeTracker.trackercond1 — Functiontrackercond1(p1, p2, delta_time, t1=(dt = (30, 100, 1300), dist=(15, 30, 120)))
Return true if the floe at p1 and the floe at p2 are within a certain distance of each other and the displacement time is within a certain range. Return false otherwise.
Arguments
p1: coordinates of floe 1p2: coordinates of floe 2delta_time: time elapsed from image day 1 to image day 2t: tuple of thresholds for elapsed time and distance
sourceIceFloeTracker.trackercond2 — Functiontrackercond2(area1, ratios, t2=(area=1200, arearatio=0.28, majaxisratio=0.10, minaxisratio=0.12, convex_area=0.14))
Set of conditions for "big" floes. Return true if the area of the floe is greater than t2.area and the similarity ratios are less than the corresponding thresholds in t2. Return false otherwise.
sourceIceFloeTracker.trackercond3 — Functiontrackercond3(area1, ratios, t3=(area=1200, arearatio=0.18, majaxisratio=0.07, minaxisratio=0.08, convex_area=0.09))
Set of conditions for "small" floes. Return true if the area of the floe is less than t3.area and the similarity ratios are less than the corresponding thresholds in t3. Return false otherwise
sourceIceFloeTracker.update! — Methodupdate!(match_total::MatchedPairs, matched_pairs::MatchedPairs)
Update match_total with the data from matched_pairs.
sourceIceFloeTracker.watershed_ice_floes — Methodwatershed_ice_floes(intermediate_segmentation_image;)
Performs image processing and watershed segmentation with intermediate files from segmentation_b.jl to further isolate ice floes, returning a binary segmentation mask indicating potential sparse boundaries of ice floes.
Arguments
-intermediate_segmentation_image: binary cloudmasked and landmasked intermediate file from segmentation B, either SegB.not_ice_bit or SegB.ice_intersect
sourceIceFloeTracker.watershed_product — Methodwatershed_product(watershed_B_ice_intersect, watershed_B_not_ice;)
Intersects the outputs of watershed segmentation on intermediate files from segmentation B, indicating potential sparse boundaries of ice floes.
Arguments
watershed_B_ice_intersect: binary segmentation mask from watershed_ice_floeswatershed_B_not_ice: binary segmentation mask from watershed_ice_floes
sourceIceFloeTracker.@persist — Macro@persist img fname
+min_area_opening::Int64=20
)
Cleans up past segmentation images with morphological operations, and applies the results of prior watershed segmentation, returning the final cleaned image for tracking with ice floes segmented and isolated.
Arguments
segmentation_B_not_ice_mask: gray image output from segmentation_b.jlsegmentation_B_ice_intersect: binary mask output from segmentation_b.jlsegmentation_B_watershed_intersect: ice pixels, output from segmentation_b.jl ice_labels: vector of pixel coordinates output from find_ice_labels.jlcloudmask.jl: bitmatrix cloudmask for region of interestlandmask.jl: bitmatrix landmask for region of interestmin_area_opening: threshold used for area opening; pixel groups greater than threshold are retained
sourceIceFloeTracker.segmented_ice_cloudmasking — Methodsegmented_ice_cloudmasking(gray_image, cloudmask, ice_labels;)
Apply cloudmask to a bitmatrix of segmented ice after kmeans clustering. Returns a bitmatrix with open water/clouds = 0, ice = 1).
Arguments
gray_image: output image from ice-water-discrimination.jl or gray ice floe leads image in segmentation_f.jlcloudmask: bitmatrix cloudmask for region of interestice_labels: vector if pixel coordinates output from find_ice_labels.jl
sourceIceFloeTracker.sort! — Methodsort!(tracked::Tracked)
Sort the floes in tracked by area in descending order.
sourceIceFloeTracker.sort_floes_by_area! — Methodsort_floes_by_area!(props)
Sort floes in props by area in descending order.
sourceIceFloeTracker.timestamp — Methodtimestamp(fname)
Attach timestamp to fname.
sourceIceFloeTracker.trackercond1 — Functiontrackercond1(p1, p2, delta_time, t1=(dt = (30, 100, 1300), dist=(15, 30, 120)))
Return true if the floe at p1 and the floe at p2 are within a certain distance of each other and the displacement time is within a certain range. Return false otherwise.
Arguments
p1: coordinates of floe 1p2: coordinates of floe 2delta_time: time elapsed from image day 1 to image day 2t: tuple of thresholds for elapsed time and distance
sourceIceFloeTracker.trackercond2 — Functiontrackercond2(area1, ratios, t2=(area=1200, arearatio=0.28, majaxisratio=0.10, minaxisratio=0.12, convex_area=0.14))
Set of conditions for "big" floes. Return true if the area of the floe is greater than t2.area and the similarity ratios are less than the corresponding thresholds in t2. Return false otherwise.
sourceIceFloeTracker.trackercond3 — Functiontrackercond3(area1, ratios, t3=(area=1200, arearatio=0.18, majaxisratio=0.07, minaxisratio=0.08, convex_area=0.09))
Set of conditions for "small" floes. Return true if the area of the floe is less than t3.area and the similarity ratios are less than the corresponding thresholds in t3. Return false otherwise
sourceIceFloeTracker.update! — Methodupdate!(match_total::MatchedPairs, matched_pairs::MatchedPairs)
Update match_total with the data from matched_pairs.
sourceIceFloeTracker.watershed_ice_floes — Methodwatershed_ice_floes(intermediate_segmentation_image;)
Performs image processing and watershed segmentation with intermediate files from segmentation_b.jl to further isolate ice floes, returning a binary segmentation mask indicating potential sparse boundaries of ice floes.
Arguments
-intermediate_segmentation_image: binary cloudmasked and landmasked intermediate file from segmentation B, either SegB.not_ice_bit or SegB.ice_intersect
sourceIceFloeTracker.watershed_product — Methodwatershed_product(watershed_B_ice_intersect, watershed_B_not_ice;)
Intersects the outputs of watershed segmentation on intermediate files from segmentation B, indicating potential sparse boundaries of ice floes.
Arguments
watershed_B_ice_intersect: binary segmentation mask from watershed_ice_floeswatershed_B_not_ice: binary segmentation mask from watershed_ice_floes
sourceIceFloeTracker.@persist — Macro@persist img fname
@persist(img,fname)
@persist img
@persist(img)
@persist img fname ts
-@persist(img, fname, ts)
Given a reference to an image object img, the macro persists (saves to a file) img to the current working directory using fname as filename. Returns img.
Arguments
img: Symbol expression representing an image object loaded in memory.fname: Optional filename for the persisted image.ts: Optional boolean to attach timestamp to fname.
sourceIceFloeTracker.MatchedPairs — TypeContainer for matched pairs of floes. props1 and props2 are dataframes with the same column names as the input dataframes. ratios is a dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios. dist is a vector of (pixel) distances between paired floes.
sourceIceFloeTracker.MatchedPairs — MethodMatchedPairs(df)
Return an object of type MatchedPairs with an empty dataframe with the same column names as df, an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios, and an empty vector for distances.
sourceIceFloeTracker.Tracked — TypeContainer for final matched pairs of floes. data is a vector of MatchedPairs objects.
sourceIndex
IceFloeTracker.MatchedPairsIceFloeTracker.MatchedPairsIceFloeTracker.TrackedBase.isequalIceFloeTracker._adjust_histogramIceFloeTracker._bin9todecIceFloeTracker._branch_filterIceFloeTracker._operator_lutIceFloeTracker.absdiffmeanratioIceFloeTracker.add_paddingIceFloeTracker.add_passtimes!IceFloeTracker.addfloemasks!IceFloeTracker.addmatch!IceFloeTracker.appendrows!IceFloeTracker.apply_cloudmaskIceFloeTracker.apply_landmaskIceFloeTracker.atan2IceFloeTracker.binarize_landmaskIceFloeTracker.branchIceFloeTracker.branch_candidates_funcIceFloeTracker.bridgeIceFloeTracker.bwareamaxfiltIceFloeTracker.bwareamaxfilt!IceFloeTracker.bwdistIceFloeTracker.bwperimIceFloeTracker.bwtraceboundaryIceFloeTracker.callmatchcorrIceFloeTracker.check_fnameIceFloeTracker.clockwiseIceFloeTracker.compute_ratiosIceFloeTracker.compute_ratios_conditionsIceFloeTracker.connected_background_countIceFloeTracker.corrIceFloeTracker.corrIceFloeTracker.counterclockwiseIceFloeTracker.create_cloudmaskIceFloeTracker.create_landmaskIceFloeTracker.cropfloeIceFloeTracker.crosscorrIceFloeTracker.deleteallbut!IceFloeTracker.detect_move!IceFloeTracker.discriminate_ice_waterIceFloeTracker.distIceFloeTracker.filt_except_labelIceFloeTracker.filt_except_label!IceFloeTracker.find_ice_labelsIceFloeTracker.find_reflectance_peaksIceFloeTracker.fixzeroindexing!IceFloeTracker.fname_ext_spliceIceFloeTracker.fname_ext_splitIceFloeTracker.from_toIceFloeTracker.get_areasIceFloeTracker.get_max_labelIceFloeTracker.getbboxcolumnsIceFloeTracker.getbestmatchdataIceFloeTracker.getcentroidIceFloeTracker.getcentroidcolumnsIceFloeTracker.getcollisionsIceFloeTracker.getcollisionslocsIceFloeTracker.getfloemasksIceFloeTracker.getidxmostminimumeverythingIceFloeTracker.getidxofrowIceFloeTracker.getpropsday1day2IceFloeTracker.getsizecomparabilityIceFloeTracker.gradIceFloeTracker.gradIceFloeTracker.hbreakIceFloeTracker.hbreak!IceFloeTracker.imextendedminIceFloeTracker.imsharpenIceFloeTracker.imsharpen_grayIceFloeTracker.isfloegoodmatchIceFloeTracker.isincountourlistIceFloeTracker.ismemberIceFloeTracker.kmeans_segmentationIceFloeTracker.loadimgIceFloeTracker.make_filenameIceFloeTracker.make_hbreak_dictIceFloeTracker.make_lutIceFloeTracker.make_psi_sIceFloeTracker.make_psi_sIceFloeTracker.makecontourstartatPIceFloeTracker.makeemptydffromIceFloeTracker.makeemptyratiosdfIceFloeTracker.matchcorrIceFloeTracker.meanIceFloeTracker.mismatchIceFloeTracker.modenanIceFloeTracker.moveIceFloeTracker.normIceFloeTracker.normalize_imageIceFloeTracker.normalizeangleIceFloeTracker.padnhoodIceFloeTracker.pairfloesIceFloeTracker.regionpropsIceFloeTracker.regionprops_tableIceFloeTracker.remove_paddingIceFloeTracker.renamecols!IceFloeTracker.resample_boundaryIceFloeTracker.roundtoint!IceFloeTracker.segmentation_AIceFloeTracker.segmentation_BIceFloeTracker.segmentation_FIceFloeTracker.segmented_ice_cloudmaskingIceFloeTracker.sort!IceFloeTracker.sort_floes_by_area!IceFloeTracker.timestampIceFloeTracker.trackercond1IceFloeTracker.trackercond2IceFloeTracker.trackercond3IceFloeTracker.update!IceFloeTracker.watershed_ice_floesIceFloeTracker.watershed_productIceFloeTracker.@persist
Settings
This document was generated with Documenter.jl version 0.27.21 on Monday 16 October 2023. Using Julia version 1.7.3.
Given a reference to an image object img, the macro persists (saves to a file) img to the current working directory using fname as filename. Returns img.
Arguments
img: Symbol expression representing an image object loaded in memory.fname: Optional filename for the persisted image.ts: Optional boolean to attach timestamp tofname.
IceFloeTracker.MatchedPairs — TypeContainer for matched pairs of floes. props1 and props2 are dataframes with the same column names as the input dataframes. ratios is a dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios. dist is a vector of (pixel) distances between paired floes.
IceFloeTracker.MatchedPairs — MethodMatchedPairs(df)Return an object of type MatchedPairs with an empty dataframe with the same column names as df, an empty dataframe with column names area, majoraxis, minoraxis, convex_area, area_under, and corr for similarity ratios, and an empty vector for distances.
IceFloeTracker.Tracked — TypeContainer for final matched pairs of floes. data is a vector of MatchedPairs objects.
Index
IceFloeTracker.MatchedPairsIceFloeTracker.MatchedPairsIceFloeTracker.TrackedBase.isequalIceFloeTracker._adjust_histogramIceFloeTracker._bin9todecIceFloeTracker._branch_filterIceFloeTracker._operator_lutIceFloeTracker.absdiffmeanratioIceFloeTracker.add_paddingIceFloeTracker.add_passtimes!IceFloeTracker.addfloemasks!IceFloeTracker.addmatch!IceFloeTracker.appendrows!IceFloeTracker.apply_cloudmaskIceFloeTracker.apply_landmaskIceFloeTracker.atan2IceFloeTracker.binarize_landmaskIceFloeTracker.branchIceFloeTracker.branch_candidates_funcIceFloeTracker.bridgeIceFloeTracker.bwareamaxfiltIceFloeTracker.bwareamaxfilt!IceFloeTracker.bwdistIceFloeTracker.bwperimIceFloeTracker.bwtraceboundaryIceFloeTracker.callmatchcorrIceFloeTracker.check_fnameIceFloeTracker.clockwiseIceFloeTracker.compute_ratiosIceFloeTracker.compute_ratios_conditionsIceFloeTracker.connected_background_countIceFloeTracker.corrIceFloeTracker.corrIceFloeTracker.counterclockwiseIceFloeTracker.create_cloudmaskIceFloeTracker.create_landmaskIceFloeTracker.cropfloeIceFloeTracker.crosscorrIceFloeTracker.deleteallbut!IceFloeTracker.detect_move!IceFloeTracker.discriminate_ice_waterIceFloeTracker.distIceFloeTracker.filt_except_labelIceFloeTracker.filt_except_label!IceFloeTracker.find_ice_labelsIceFloeTracker.find_reflectance_peaksIceFloeTracker.fixzeroindexing!IceFloeTracker.fname_ext_spliceIceFloeTracker.fname_ext_splitIceFloeTracker.from_toIceFloeTracker.get_areasIceFloeTracker.get_max_labelIceFloeTracker.getbboxcolumnsIceFloeTracker.getbestmatchdataIceFloeTracker.getcentroidIceFloeTracker.getcentroidcolumnsIceFloeTracker.getcollisionsIceFloeTracker.getcollisionslocsIceFloeTracker.getfloemasksIceFloeTracker.getidxmostminimumeverythingIceFloeTracker.getidxofrowIceFloeTracker.getpropsday1day2IceFloeTracker.getsizecomparabilityIceFloeTracker.gradIceFloeTracker.gradIceFloeTracker.hbreakIceFloeTracker.hbreak!IceFloeTracker.imextendedminIceFloeTracker.imsharpenIceFloeTracker.imsharpen_grayIceFloeTracker.isfloegoodmatchIceFloeTracker.isincountourlistIceFloeTracker.ismemberIceFloeTracker.kmeans_segmentationIceFloeTracker.loadimgIceFloeTracker.make_filenameIceFloeTracker.make_hbreak_dictIceFloeTracker.make_lutIceFloeTracker.make_psi_sIceFloeTracker.make_psi_sIceFloeTracker.makecontourstartatPIceFloeTracker.makeemptydffromIceFloeTracker.makeemptyratiosdfIceFloeTracker.matchcorrIceFloeTracker.meanIceFloeTracker.mismatchIceFloeTracker.modenanIceFloeTracker.moveIceFloeTracker.normIceFloeTracker.normalize_imageIceFloeTracker.normalizeangleIceFloeTracker.padnhoodIceFloeTracker.pairfloesIceFloeTracker.regionpropsIceFloeTracker.regionprops_tableIceFloeTracker.remove_paddingIceFloeTracker.renamecols!IceFloeTracker.resample_boundaryIceFloeTracker.roundtoint!IceFloeTracker.segmentation_AIceFloeTracker.segmentation_BIceFloeTracker.segmentation_FIceFloeTracker.segmented_ice_cloudmaskingIceFloeTracker.sort!IceFloeTracker.sort_floes_by_area!IceFloeTracker.timestampIceFloeTracker.trackercond1IceFloeTracker.trackercond2IceFloeTracker.trackercond3IceFloeTracker.update!IceFloeTracker.watershed_ice_floesIceFloeTracker.watershed_productIceFloeTracker.@persist