After some time away due to a new job & the NASA Space Apps Challenge hackathon, I was finally able to sit down for a day of debugging my fork of SimpleCV. I left off with a version that loads and functions on frozen-in-time dependencies on my Raspberry Pi, but that setup left me with no real-time image capture capabilities. The camera I run on my Pi is a discount Arduino camera, which may explain why SimpleCV was not able to recognize the device driver. It works from the command line, but not within the SimpleCV shell.
Luckily, I was use my progress on the Pi to get my big-girl computer to the point that it can now load SimpleCV in the Python 2.7 virtual environment I have for working on it! Since I already have a version that works on historical versions of dependencies, I’m ready to start moving the internals forward in time. The biggest issue is changing the processing functions from OpenCV 2.x to 3.x, especially since the removal of cv2.cv came with significant changes in names and parameters of image processing functions.
Taking a step to the side, then, I’m working through OpenCV directly. This documentation tutorial on changing color-spaces was one such exercise.
So far, it is looking like many of the changes from cv2.cv to cv2 are more than just different function/method/parameter names, though some fixes have been as simple as removing ‘CV_’ from certain constants. However, the more familiar I become with these libraries, the better my attempts at addressing the endless tracebacks will be. I know this is more than a 1-person workload, but I still feel like updating this library is a valuable project and an attainable goal.