Study On Auto-detect System For Detecting Defects Of Workpieces' Electroplated Surface

Quality is one of the essential factors that affect the survival and development of an enterprise. For this purpose, one important means is to automatically inspect material and product during the process of produce and feedback the inspection information into quality control. What this thesis being studied is the stochastic defects on the surface of electroplated workpieces. We expect to make out all the defective workpieces and take them off and let the quality detection come true. The reflection of the electroplated workpieces surface is very strong. Automatic detection for defects on strongly reflective surface is an important research thesis, which is not easy to solve too. This thesis have been studying for quite a long time. When studying the relation of the illumination of workpiece surface and the difficulty to detect defects, the concept of defects' expressivity was brought out. My former had also studied the illumination system and many methods in image processing. This paper continue studying on the illuming system and image processing methods and make out all detection methods for each kind of defects. On the shoulder of my former, combined with CCD photoelectricity inspection and image processing and automatic control technique, This paper presented out an methods and auto-detect system for detecting the defects of workpieces' electroplated surface. The main content of this thesis are:1.In research of the character of the defects of workpieces' electroplated surface I found that the Defect Expressivity is related to not only the illumination system but also the distance from workpieces to lamp-house and CCD.2.After studied and designed several kind of lamp-house I finally designed a conoid LED illumination system with uniform scatter light, which provides a good illumination condition for CCD and helps improving the quality of images.3.The pilot study on disparting machine and overturn machine.4.Putting forward different detecting methods for different defects. Based on experimentation I have studied the image binaryzation arithmetic and solved the arithmetic of getting rid of white circle and circularity judging and line's max length judging. 5.Analyzing the vast experimental data and proposing the betterment suggestion.