Research On Image Defect Detection Algorithm Based On CUSUM

Currently, the quality detection of chopsticks in our country mainly rely on traditional artificial defect detection. However, there are some disadvantages of artificial detection methods, such as low efficiency, high cost, poor working environment and so on. Along with the higher production efficiency and stricter quality requirements, traditional artificial defect detection method could not guarantee the production demands. This paper researches the status of the chopsticks surface defect detection technology and the main research contents of this paper are as follows:?Aimed at chopsticks image, this paper does researches on image denoising, enhancement and edge detection to realize pre-processing of chopsticks` images.?For chopsticks images, an image segmentation method is designed. Through analyzing the percentage and percentage variance of RGB three channel pixel data, the effectiveness of the segmentation method is verified. Experimental results show that the method designed in this paper can effectively separate chopstick image from the background image, and get accurate chopsticks RGB data. The data analysis results have a certain reference value in chopsticks defect detection.?The improved CUSUM detection algorithm is applied to detect the chopsticks' defects. By analyzing the chopsticks' RGB data, we find the different characteristics between defective chopsticks and standard chopsticks.