Today I finished up the coding for the detection based on correct region. I tested it and apparently only 2 out of 9 detections were accepted. I investigated this by viewing the images detected in HSV view and realised that although some of the detections identified was in the correct area of detection, but it indirectly got dismissed as it was apart by just 5 to 7 pixels.
I then modified my code to accept the detection if the area that is detected is less than 20 pixels away from the correct detection area. I have just finished the code for this and will test this out tomorrow.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment