Corrugated Pipe Defect Detection Based on Digital Image Processing

With the increase of human cost and the shortcomings of human eye to detect the product quality, such as visual fatigue and judging standard, these factors greatly restrict the development of the quartz rod manufacturing industry. In this paper, a method of detecting quartz rod bubbles and grading their quality based on machine vision is proposed. The main research contents are as follows: 1)image preprocessing, which includes image denoising, threshold segmentation and image binarization; 2) The algorithm of the connected domain tagging based on traveling; 3)Shan bubble feature extraction, feature parameters are limited to the area, position, number of bubbles, and determine the quality grade of quartz rods according to the characteristic parameters. Shan According to the algorithm presented in this paper, 100 samples are tested, and the correct rate is 96%, The improved position adaptive threshold segmentation algorithm proposed in this paper is better than the Ostu algorithm for the whole two-dimensional image. Compared with the general connected area labeling method, the run-based connected area labeling algorithm proposed in this paper consumes less time and has higher accuracy.