SwapnilBhole 1: I have below ok image. and below is not ok image. What are algorithm using image processing or deep learning I can try on this, so that I can make not ok image as defected image. 2: same part from backside. ok image nok ok image
jakaskerl Hi SwapnilBhole Again, the same as https://discuss.luxonis.com/d/2624-crack-detection-using-oak-d-pro-poe/2. You should be able to train a model that recognizes both cracks as well as discoloration as a defect. Thanks, Jaka
SwapnilBhole @jakaskerl Is there any algorithm using image processing that can useful for discoloration as a defect. I will try using deep learning, but, just wanted to know, how this can be done using image processing.
jakaskerl SwapnilBhole Thresholding Color Thresholding: If the discoloration is of a specific color, you can isolate that color using color thresholding. For example, if the defect is a specific shade of red, you can set thresholds to capture just those red tones. Adaptive Thresholding: Useful when lighting conditions vary across different parts of the image. Histogram Analysis Analyze the color histogram of the image to identify significant deviations in color distribution. This can be effective if the discoloration affects the color distribution noticeably. Clustering Techniques like K-means clustering can be used to segment the image based on color or intensity. This can help in isolating regions with discoloration. Thanks, Jaka
