Skip to content

NOTEPAD of tips to include

Shawn Harrison edited this page Dec 29, 2016 · 2 revisions

As I attempt to build my own 'home-made' Argus system, I often seek advice from John Stanley (who has a wealth of information from decades of experience). I aim to incorporate those tips into the wiki somewhere. However, until those tips find their way into the appropriate location, they will temporarily reside in this notepad.

Camera Settings:

It is important will make your life easier to ensure that the extrinsic parameters (e.g. camera geometry) remain constant throughout image capture activities. Obviously the camera position changes constantly when observing from a UAV platform, but post-processing requires greater effort to resolve the camera position. To help minimize changes to extrinsic parameters, use a fixed-focal length lens with fixed aperture. A constant aperture is achieved by fixing the iris (the diaphragm within the lens), typically with a locking knob to prevent accidental adjustments. John suggests fixing the iris at f8. He lets the capture software automatically adjust exposure time and signal gain (within pre-specified limits) in order to improve appearance of the image digitally rather than by physically adjusting the iris. Digital changes to the image capture do not impact the camera geometry. As things get darker (e.g. during overcast or low sun-angle) John lengthens the exposure until it hits the upper limit previously set (usually 0.1s) and then increases gain. Increasing gain increases noise, so it should not be done without specific need. Increasing exposure also decreases the image sharpness of moving targets (e.g. wave crests) - hence a relatively fast upper limit.

Image Capture:

There are many ways to control cameras in order to capture images. Many cameras come with proprietary API languages (e.g. Point Grey Blackfly cameras have FlyCapture2). However, depending on the type of camera interface (USB3, Firewire, GigE) there are also international standards established common for all camera makes. For example, GigE Vision is an interface standard and all cameras supporting GigE Vision interact the same way with software also supporting GigE Vision. The standard defines required elements for camera identification, control, and output. It uses Generic Interface for Cameras ('GenICam'), a programming interface for camera attribute control.

John Stanley uses a program that he wrote called HOTM. Some people use Streams7, a commercially-available control software from I/O Industries. And some people write project-specific capture routines with the camera API.

  1. Mounting: This camera is so tiny and the lens is much heavier. I feel like it would be better to mount the lens to something using one of those collar mounts rather than the little tripod adapter mount on the bottom of the camera housing.

I've had no problem using the camera mounting adapter even with larger lenses.

Eventually I imagine stuffing the unit in a Pelco EH2508 housing (or similar). In your experience are those cyclindrical housings ok or do you prefer the rectangular ones?

I currently use the cylindrical ones. There is no significant difference in the shape, but there is a problem you'll see in the long run with a flip-top housing. The body of the housing is aluminum, the latch is stainless steel, and the pop rivets holding the latch to the housing are regular steel The rivets will corrode. You'll come back to your station and see that the top has flipped up with the latch hanging off the back end of the top. A cable tie solves this, but is not elegant.

Do they have a 1.4"-20 nut inside like a typical tripod, or how is the camera/lens mounted inside?

They have a long slot or multiple holes that a 1/4" machine screw will fit through.

Clone this wiki locally