Hi Forum, I'm new to OpenCV and image processing world. The company I work at wants to employ computer vision to identify the box pallet configuration correctness. My initial thought is capturing images of front, back, right, left and top sides of a pallet on top of conveyor . From right, left and top I can get perfect rectangle view of boxes on top of pallet. From front and back I will get a little bit warped images because of the camera positions.
To give you the idea, I took a warped picture of a pallet with boxes on top of it, crop and warp it. Sorry I blurred the image to hide the brand. Manually I can determine the four corner positions, crop and warpPerspective using the information. Is it possible to do it automatically?
Original:
Cropped and Warped:
Thank you very much.