The Research And Application Of Defects Intelligent Detection Technology Based On Image Processing Technology

With the rapid development of computer technology,the computer aide detection has become an important direction of this area. At present, the detection of aluminum wheel is normaly completed by human, and the efficiency is very low. This paper has developed a set of aluminum wheels defects intelligent detection system, which contains the basic functions of the image processing, and it can detecs the dafect and the number of intelligent detection s; it also can carry out the characteristic parameters of defects, and give the defect level judge.First of all, according to request ,it give a hardware system. In the design of software ,it give the five software modules: image acquisition module, database modules, image processing and analysis module, defects extraction module, defect analysis module. First in image acquisition, the image can be deposited directly into the database directly, Database using SQL Server 2000 and the ADO interface to design. Database can get a large amount of storage. Image processing includes: image noise reduction, the gray-scale transformation, image sharpening, edge detection. For the image noise, according to the character, using a multi-frame superimposed average and median filtering method for processing. By improving the median filtering, it get a good effect; for the low contrast,using the linear and the logarithm transformation, the image contrast is enhanced; for the fuzzy image, according to its characteristic, what is more, after the analyzing of the image, it is helpful to use gradient peaking based on the judgment of threshold. This paper shows a kind of edge detection based on the brink of SUSAN ,after analysising the traditional ones, and it get the desired results. In the defects extraction part, effect is not good by division the image directly. After analyzing the reasons ,it give the defect extraction on the basis of ROI and based on the morphology. The defect detection mainly bases on the marking method. It chose perimeter, area, circular and center of gravity as the characteristic parameters of defects. the defect level is determined ,on the basis of casting defects leverl standard ,so the system can determine the level of the defect. Finally, it compares the defect detection based on ROI with the one based on the morphology, and it also give the applicable scope of the two ones.