I am having some challenging problem with trying to perform IC surface defects detection using Python OpenCV.
I have problem with ignoring the legs and the circle that is naturally on the IC surface. Which for me is almost impossible. Could you suggest some solutions ? (No AI or Deep Learning)
Take a reference image of the object that is clean. Then warp one of the two images to the same orientation and position and scale. Then get the difference between the two images. If the difference is low, it has few or no scratches. If the difference is high it has scratches. Also the difference image will be dark if no scratches and will show the scratches if the object is defective.
Take a reference image of the object that is clean. Then warp one of the two images to the same orientation and position and scale. Then get the difference between the two images. If the difference is low, it has few or no scratches. If the difference is high it has scratches. Also the difference image will be dark if no scratches and will show the scratches if the object is defective.