Study On Roundness Image Testing Technology And Application

Computer vision through transforming the images which captured by camera into pixel, and based on the brightness and color of pixel, the information transformed into digital signal, which transmitted to image processing system, then using digital image processing technology to process and analyze the image, finally realized the control of the equipment based on the analyze result. Using computer vision technology to deal with image measurement is one of the applications of computer vision. Image measurement includes image capture, image preprocessing, feature extraction and size measurement. The accuracy of each part will affect the accuracy of final measurement. Therefore, it is necessary to research and optimize every single part of image measurement.The main research of this paper is roundness measurement technology of image based on computer vision. Chose badminton as investigated subject, discussed and realized the applications of image measurement technology on edge and roundness measurement of badminton. The sequence image of badminton rotating in wind tunnel was captured by camera. Image processing technology was used to analyze the information of image sequence, and finally realized the on line none-contact measurement of edge and roundness of badminton. The main work of this paper as followed:1. Test system structure and Image preprocessing at multiple targets situation. Introduced the hardware and software processes of test system. Considering the complicated background and2targets, calibration method was used to locate the targets and segmented the targets. After compared several kinds of image enhancement and binaryzation methods, an improved contrast enhancement method and block dynamic threshold method have been raised. Finally got the clear and high quality Binary image.2. Image edge correction. Considering the image distortion caused by the changed relative position between badminton and camera, an image correction method raised in this paper. Common attitude angle measurement method was introduced in this paper, and a new attitude angle estimated method was raised based on the change of sequence images and priori knowledge. On the basis of the coordinates of image edge, through formula derivation and coordinate conversion, correction of image edge can be realized. This provided a high quality image for followed size measurement and roundness error analysis.3. Characteristic point locate, fit and dynamic calibration compensation of diameter. The locate accuracy of characteristic point directly affect the accuracy of size measurement of the system. A locate method was raised in this paper which is high accuracy, short operation time and convenient for programming. Least square method was used to fit all characteristic points, and then get the diameter of the circle as the diameter of badminton. A dynamic calibration compensation based on basic theory of image formulation was used to compensate the diameter of badminton, which realized online dynamic size measurement.4. Data and error analyzing. To check the feasibility of this measurement system, a lot of experiment has been done. The accuracy of badminton size measurement and the roundness error of badminton have been estimated. Finally, the error source of the system and its effect has been analyzed. Testing results shows that, the accuracy of this measurement system meets the requirement.The image preprocessing, characteristic point locates and fits algorithms programs were realized based on VC and Matlab. The measurement system has already been built. The work of this paper has some theory significance and high practical value.