Things changed up with my senior design project, and we are no longer using the $50,000 camera. With that said, we are now using a Pixy2 camera and we are sorting the marker bands by color.
Up to now, I have managed to get the Pixy2 camera working to detect colors on objects. For example, I was able to get the camera to detect the red colors on this lotion bottle (see picture below)
But I haven’t done it on objects that are as small as marker bands.
To get an idea of how small these bands are, here are some dimensions from our Requirements Document:
- Outer Diameter Range: 0.070” – 0.170”
- Wall Thickness Range: 0.0018” – 0.003”
- Length Range: 0.020” – 0.125”
FYI we painted the marker bands with some Crayola child-friendly paint that washes off with soap and water.
When the Pixy2 camera works, it works well. The image below demonstrates that (although, one green marker band is not detecting).
But it’s not consistent. Sometimes, the Pixy2 camera detects false-positives. In addition, the glare of delrin disk (the black disk where the marker bands are resting on) interferes with the detection.
I also think the marker bands are just too darn small to be detected. The picture below shows the marker bands and the paint bottles (note: this image was taken through a magnifying lamp in an attempt to make the marker bands appear larger to the camera).
PixyMon (the Pixy2 camera software) is only detecting the paint bottles.
Thus far, Theresa (one of my senior design teammates) and I have tried various camera angles and placements. The same issues of inconsistent readings and glare keep arising.
I think another issue is the room lighting. I’m noticing that in the later afternoon, the Pixy2 camera is more inconsistent.
After doing some more research, I think the next thing we have to do is to enable “Overexposure Highlighting” in PixyMon. We’ll be working on it some more today!