This week, I got the blob detectors for the bok choy and the beef running (shown below). There are still slight adjustments needed for false positives on hands, since some skin color is very similar to the color of the fatty streaks in the beef. Unfortunately, the work I did in the first half of the week was on trying to track the hot pot and strainers, and this will not be necessary going forward due to our change of plans with the control tower and touch screen. However, I’m very glad about this change overall because it will help our computer vision algorithm run faster since it won’t be running an ellipse detector every frame.
Since I was visiting NJ for the long weekend, I also picked up a 3D printer that Shane built a few years ago in order to bring it back to Pittsburgh. Hopefully our team can use it to do material-consuming prints at a fraction of the cost of using TechSpark. I also attempted to set up the Jetson Nano in order to get footage with our IMX camera, but for some reason its SD card image repeatedly failed to validate after flashing. Each attempt was a fairly time-consuming process, so I haven’t gotten a chance to thoroughly troubleshoot it yet.
This coming week, I hope to pinpoint the correct HSV range for the beef while excluding skin tones and also distinguishing from brown shiitake mushrooms. I also hope to figure out what is going wrong with the etching or flash the SD card image through a different means in order to get the Jetson Nano working.
