A fast defect detection algorithm for glass tube based on ROI reduction

n this paper, we present an algorithm for defect detection in glass products that allows us to minimize the processing time. The main idea is based on the reduction of the size of the image area to investigate by using the features of glass images. The proposed solution doesn’t compromise the quality of detection and allows us to achieve a performance gain of 66% in terms of processing time, and 3 times in term of throughput (frames per second), in comparison with standard algorithms for defect detection.