This week, I finished downscaling an image to a 160×120 pixel image by using the Pillow library in Python. I researched on the OpenCV algorithms for color tracking, and I broke the library function calls to for loops on a 2D array of pixels (for future conversion to HDL). I implemented the conversion from RGB to HSV. I also collected the mask of the pixels that fit in the HSV range of a certain color. Then, it is put through erosion and then a dilation in order to get rid of noise. Then, using the algorithm to find the largest rectangle in a binary matrix, I found the center of the tracker which will be the reference for the joint of the posture analysis.
I am currently on par with the schedule. My plan for next week is to test the algorithm on the trackers that we bought from Amazon, and to determine the lower and upper bound of the HSV in order to recognize the trackers. Also, I would start researching on the Image Posture Analysis, and help Venkata with the hardware communication protocol.