Michael’s Status Report for 2/10/24

Weekly Accomplishments:

1. Dug into OpenCV and was able to extract QR code data from a smart label
   • Ran into an issue while trying to test directly underneath a light fixture where the camera couldn’t see the qr code.
     • Issue likely is due to glare. Software attempts to reduce glare were expensive and drastically reduced frame rate on my laptop.
   • Additionally started to implement object tracking (how the glare issue was found)
2. Requested 3 Raspberry Pi Zero 2 Ws + Raspberry Pi Camera
   • Upon arrival, will be used to test whether or not RPi Zero is powerful enough for image capture + streaming
3. Started research into fruit recognition models
   1. List of public github projects related to fruit/vegetable classification
   2. Further research needs to be done into models especially for angled + lower resolution views of produce.
   3. Found labeled dataset with fruits & vegetables if training own model

Progress

• (Currently non-blocking) More work needs to be done to make the smart label recognition more robust (i.e. object tracking,  glare issue)
  • In meeting on Wednesday, pick Professor Savvides’s brain to see if polarization lens are a good idea.
  • Delay working more on smart labels until after design presentation.
• (Currently non-blocking) Explore accuracy of fruit/vegetable models especially at non-ideal angles.
  • Take pictures during next trip to Giant Eagle
  • Test models against real test images

Next Week’s Goals

• Work w/ Professor Bain to get access to AMD’s Kria KV260 platform
• Decide on hardware for scanner iteration 1
• Create CAD model of scanner
• Prepare for design presentation (slides + presentation prep)