The Study Of Measuring The Dimension Of Shaft-type Elements By The Machine Vision

In the mechanical industry, the shaft-type elements are an important part of the machines. With the unceasing improvement of the quality of machine part, the method of the evaluation and the measure to the dimension of the part sets a higher request.At present, the contact dimension measurement applies more widely in the mechanical industry; its characteristic is intuitionist, reliable and easy for operation. It can satisfy the need of the general measure. But because the contact type measure technique has its certain limitation: 1) on some bad circumstance or dynamic condition, the contact type measure technique is unfit. 2) The tip contacts the object surface direct, so it is easy to scratch the surface. 3) It is unfit for the on-line and real-time inspection. Therefore, people have launched extensive research to each non- contact type measuring technique. In recent years, along with the rapid development of photo electricity technology and the computer technology, the measuring technique of the dimension of mechanical part based on the Machine Vision technology receives many academicians'more attention. It has the characteristic of non- contact, fast, convenient and powerful in data processing etc.In this paper, it has researched and studied the Machine Vision and image processing technology, then designed and built the hardware of measure system. Using the CCD to capture the image of the axes, and then this primitive image is send to the computer image processing system after digitized to carry on next step of processing. First, the hardware of measure system should be calibrated, and this is the first-line work. Then, it is the process of Image processing, and this is the most important and pivotal work. In this process, the Techniques of Super-Resolution Image restoration is used to improve the resolution of the image and the precision of the measure. Then, detecting the edge of the axes in the image based on the method of Sub-Pixel, and at last the result can be obtained by computing. 1. In the measure of the dimension of the machine part, it only needs to calibrate the distance between two points of the part which corresponding pixel in the image. Therefore, it has presented a new method in this paper. It can obtain a model with optimize method by experiments. Using the model to calibrate the system is very easy , and the result is very good. The distance between the axes and the camera lens and the distance between the pixel and the center of the image are confirmed in the model. Therefore, the model is be becoming to calibrate the measure system in fieldwork.2. The Techniques of Super-Resolution Image restoration is used to improve the resolution of the original image and the precision of the measure firstly. At first, it chooses some images to obtain the image which are more pixels than before by method of interpolation. Then the Motion Estimation Techniques is used to get the motion vectors and select the four original images which are appropriate. At last, the four images are used to construct the perfect image which is clearer.3. The edge detecting is an very important process of the measure. It chooses the method of Sub-Pixel to detect the edge of the axes in the image in this paper. In the process, the rough edge of the axes should be found at first and then detect the edge accurately using the Gray moment operators. It chooses five pixels which in the area of the edge for the characteristic of the detecting method, and the edge between the two pixel in the center of the five.This paper is composed of seven chapters. The first chapter introduced the background and the significance of this work, and summarized several existing dimension of axes measuring technique, finally elaborated the main content of this paper; The second chapter has made a detailed introduction for the Machine Vision and digital image processing technology; The third chapter has made a brief introduction for the measure system first, and then introduced calibration of the system detailed. The fourth chapter has explained the Techniques of Super-Resolution Image restoration which had been used to improve the resolution of the original image and the precision of the measure. The fifth chapter introduced the edge detecting method at present first and then explained the method of Sub-Pixel for edge detecting which had been applied in this paper detailed. The sixth chapter introduced the hardware of the measure system first and then explained the process of the measure detailed, and at last made an analysis of the precision of the measure. The seventh chapter is the part of conclusion, it summarize the production of this paper and point out the insufficiencies of this paper at last.