Image Stereo-Pairs

Problem we are trying to solve

Project Description - "Create a software system that produces an image stereo-pair from a single, ortho-corrected image and a height (depth) field with identical sample spacing. Ortho-corrected satellite imagery and the corresponding pixel-for-pixel height fields are readily available and are used as inputs to scientific models. We seek a process that will create an image stereo-pair for visualization purposes, utilizing the two readily available sources above. We are interested in software to both create and validate the process."

We - Abemelech, Deepak, Sofonias, and Soobin (ADSS) - are planning on creating an interface for users to input Ortho-corrected satellite imagery with depth field data and output stereo-pair images.

Installation

To use this software, clone this repository and install the required dependencies:

For Windows

git clone https://github.com/your-repository/stereo-depth.git
cd stereo-depth
python -m venv venv
venv/Script/activate 
pip install -r requirements.txt

For Linux/Mac

git clone https://github.com/your-repository/stereo-depth.git
cd stereo-depth
python -m venv venv
source venv/bin/activate 
pip install -r requirements.txt

Available Parameter

usage: main.py [-h] -l LEFT -r RIGHT -b BASELINE -f FOCAL [-d DEPTH] [--validate] [--execute]

options:
  -h, --help            show this help message and exit
  -l LEFT, --left LEFT  The Path of left stero image
  -r RIGHT, --right RIGHT
                        The Path of right stero image
  -b BASELINE, --baseline BASELINE
                        The baseline of the stereo image
  -f FOCAL, --focal FOCAL
                        The focal-length of the stereo image in pixels
  -d DEPTH, --depth DEPTH
                        The Path of the depth image (only required for validation)
  --validate            Validate the depth map
  --execute             Execute the disparity and depth map processing

Usage

To use the software, provide the left and right stereo images as inputs. The software supports two main operations: --execute for processing and generating depth maps, and --validate for validating generated maps against provided depth maps.

Execution

python main.py --execute -l path/to/left_image.jpg -r path/to/right_image.jpg -b <baseline> -f <focal-length pixel>

This command processes the provided stereo images to generate a depth map. The results will be saved as 'test.png'.

NOTE: We have example images in the images folder for testing purposes (Baseline = 0.6, Focal-length in pixel = 645.24)

Validation

python main.py --validate -l path/to/left_image.jpg -r path/to/right_image.jpg  -b <baseline> -f <focal-length pixel> -d path/to/provided_depth_map.csv

This command validates the generated depth map against the provided reference depth map. The accuracy metrics will be displayed in the console.