The Saw Defect Detection Based On Machine Vision And System Development

The saw widely use in the industry manufacture area, mostly export to Europe and American, and it maily has these character:strong solidity and flexi-bility, anti-heating, anti-abrasion. The surface of saw spray white oil painting, printing logo、character、code and so on. In the process of printing, it may make the wrong logo or scratch the logo. The traditional testing method dependon the worker’s eye-measurement, due to the large production quanlity, the worker has strong subjectivity and easy lead to visual fatigue, this method will lead to wrong judgement and affect the efficiency. But the machine vision possess the high efficiency and persistence, can work under the complex environment, so it is better than the traditional testing method. Because of it can strongly improve the manufactere efficiency, it already widely uses in every area. So this paper combine with practical using, research the saw defect detection based on machine vision and system, testing the faultiness of saw more effective and faster.The paper firstly introduce the original of machine vision, the development situation at home and abroad and the applications at painting industry. According the production environment and the client’s requirement, to make sure the machine vision hardware. According the feature of saw surface flaw, constitute the systerm of machine vision image preprocessing method and testing the saw surface flaw, then research the algorithm deeply. Lastly combine the software and hardware to build up the saw defect detection systerm base on the machine vision.The main contents of the paper:according to the pratical production, testing 4 pieces saw, base on contour perimeter to get the target. In the light of the limit of testing assembly line, calculating the saw deviation angle, base on the affine transformation to correct the saw’s horizontal level. Using vertical projection to divide the saw into 8 sub-element, then counting the height and length and base on the template matching to locate the target, using the prinicple of different frame method, the target minus the model image, lastly make the output image binary conversion, this method can quickly and exactly test the saw’s surface flaw. According to the density of teeth, it can use this method to testing:pick up one equal length saw, base on the pixel traversal to counting the teeth number. On the basis of the production testing requirement, choose the appropriate camera、shot and light source. Using Visual Studio 2010 to the developing tool, base on the Open Source Computer Vision Library to develop the machine vision testing systerm.Through the testing out come and the analyse experience, the machine vision saw surface flaw testing systerm have favorable robustness, provide technical guidance for the factory automation testing.