Research On The Machine Vision Detection Of Workpiece Surface Crack

Surface crack detection of workpiece is an important part of industrial surface quality detection. In people's daily life and production, the quality of workpiece is directly related to personal safety and property safety, and surface quality is an important factor affecting holistic quality, so the detection of workpiece surface crack is indispensability.Traditional method of workpiece surface detection relies on human eyes. Because of the weaknesses of visual fatigue, poor environment, poor real-time performance problems by artificial detection in modern high-speed production-line, it is difficult to guarantee the reliability and sustainability. In recent years, benefits from rapid development of computer technology, photoelectric technology and image processing technology, machine vision receives great attention and development in industrial surface detection, automated assembly equipment, traffic monitoring and other fields. Compared with the traditional artificial human visual inspection, machine vision has a series of advantages such as high efficiency, high precision and multifunction.The detection target is workpiece surface crack in this paper. By fully analyzing the characteristic of surface crack image, this paper systematically studies image processing algorithms and software implementation, so as to using machine vision replacing tradition artificial detection.This paper firstly introduces the basic theory of the machine vision technology, the development and research situation at home and abroad and the application in industrial surface detection. Based on the analysis of characteristic of workpiece and general situation of crack existing, this paper then sets up the system hardware platform, assesses software solution and formulates overall sys-tem detection scheme.The workpiece surface image processing is the key of machine vision. According to the basic process of image processing, this paper successively studies a series of commonly used algorithms referring to gray method, de-noising method, enhancement method and segmentation algorithm and so on, and symbolically puts forward the block adaptive fuzzy enhancement algorithm to enhance the contrast of crack region an image. Besides, this paper also analyzes the different de-noising results by different parameters of wavelet threshold de-noising method.This paper then completes software design of surface crack detection based on LabVIEW software platform and NI vision software package. Some algorithms which tested effectively are applied to achieve modularization programming of image normalization, ROI extraction, pre-processing, segmentation and recognition, completing the judgement of existence of crack.In the end, this paper selectes a series of sample workpiece to test crack detection effect. The test result indicates that the detection of crack based on LabVIEW machine vision has better precision, accuracy and stability, proving crack detection system is reliable and effective, which has potential application prospect.