Share Practical Insights - Application of machine vision scratch detection

Scratch detection is a common problem in industrial production. In industrial production, many parts of equipment are working in high temperature and high-pressure environment, the load is complex, the use of poor environment, high failure rate. Therefore, the visual detection of scratches and cracks of relevant parts is particularly important.

Industrial products are very different, and the scratch detection methods and requirements are also varied. The use of machine vision for scratch detection can be basically divided into two steps:

(1) Detect whether there are scratches on the surface of the product.

(2) After determining the existence of scratches on the analyzed image, the scratches are extracted.

Despite the diversity of images in industrial production, for each image, it is necessary to undergo comprehensive analysis in order to achieve the desired effect. No matter it is scratched or pitted defect target, the defect target occupies a small pixel in the whole image, and the difference between the gray value of the defect target and the uneven background is not obvious in the area of low SNR. Therefore, in order to detect and identify defects accurately, an effective algorithm must be used to highlight the basic features of defects in the image and weaken the basic features of the background of the image.

Machine vision, as one of the important frontier technologies of artificial intelligence manufacturing, has unique advantages in hardware and plastic defect detection. Machine vision inspection technology for hardware defect detection has the advantages of non-contact, high efficiency, low cost, high degree of automation, and has great value in detecting sheet metal defects and preventing defective sheet metal products.