基于机器视觉的方块地毯边缘缺陷检测技术

Edge defect detection technology of square carpet based on machine vision

目前方块地毯生产过程中的质量检测主要由人工完成,费时费力,缺乏科学性。本文介绍了一种基于机器视觉的方块地毯边缘缺陷的自动检测技术。利用地毯边缘的特性,将图像滤除和边缘检测结合在一起,相比传统方法减少一次形态学基本运算,应用霍夫变换检测出方块地毯的四条边,将二值化图像与该四边形的差以及该四边形的正方形程度作为缺陷判别依据。实验结果表明提出的方法检测率达到92.6%。该方法具有有效性和可行性,可以达到企业的生产要求。

At present defect of square carpet are detected mainly by manual operation,which is timeconsuming and lack of scientific.This paper introduces a kind of automatic detection technology of the edge defect of square carpet based on machine vision.Due to the characteristics of the edge of the carpet,by combining image filtering and edge detection,one basic operation of morphology is reduced when compared with the traditional method.Hough Transform is used to detect the four edges of the square carpet.Using the square degree of the quadrilateral,and the difference between the binary image and the quadrilateral as the basis of defects.Experiments show that the detection rate of this method reaches92.6%.This approach is effective and feasible,and meets the production requirements of enterprises.