I am the lead programmer for Team Tempest, and we are trying to get the camera to recognize the vision targets on the goals. The camera will recognize the tape, but not create a complete goal. Can anybody help us out?
I’m assuming that you mean that part of the target box is missing in the image…
Do you have your light source close to the camera? (it should be) Ideally you’d use an LED light ring around the camera. That should provide a very clean reflected image of the target.
Are you using the NI Vision Assistant to capture and test the image and various filters? It makes it very easy to capture images and test out different filters.
Recommend looking at Miss Daisy’s code from last year.
This is my second year doing vision programming and last year we won an award at the queen city regional for our use of it and how we incorporated the Microsoft kinect with it. To track an image, first you must acquire it, then process it to eliminate noise. I use the OpenCV libraries developed by intel. It is perfect for these types of projects. My method also includes the distance in x, y, and z coordinates and rotation in pitch roll and yaw. If you are not getting a solid square back, try adding more light into the system, either ir or green i find works best, not red or blue due to the nature of the field. Best of luck from team 1706!