This past week started with writing the design review slides, and then some preparation to present in the early portion of the week (and also, a midterm). As such, I didn’t begin working on code until the end of the week. I have figured out how to run a variety of OpenCV built-in tracking algorithms on my own computer, on live video feed, with a hardcoded bounding box. See example here!
With regards to the CV coding side, I would say that I’m on schedule. With regards to the hardware rPi, I’m a bit behind, as we had hoped to order it by the end of this past week (but delayed, due to the ongoing debate about rPi versus Jetson). However, this doesn’t have any impact on our ability to draft the software, so I’m not overly concerned about the delay. The plan at this point is to get the rPi this week, before spring break. If we are unable to test it out before spring break, or have trouble working with it in initial tests, then this is something I’ll invest more time into over the break. This should put me comfortably back on schedule.
In terms of deliverables for next, first and foremost I plan to make sure we order the parts we need, and to begin to poke around with using them. For the OpenCV stuff, I have 3 goals: 1) get rid of the hardcoded bounding box in favor of determining them dynamically, via something like pixel differences; 2) explore/experiment a bit more to determine which tracking algorithm seems best for our project; and 3) figure out how it will interface with the ML side (how to determine when to request ID, and how to cutout the important part of the image).