The Study On Template-Based Entire Imaging Measuring Method For Microparts

With the development of computer techniques, image measuring techniques is more and more widely used in the inspection of industrial parts. However, in the image measuring of microparts, the imaging area of the measuring system needs to be reduced to obtain high measurement precision. When numbers of points on micropart are measured, the frequent relative movements between the imaging system and the micropart are needed. So many mechanism movements are required to measure just one micropart. The efficiency of the measurement is hard to be improved in this way. And it is also difficult to achieve automatic and intelligentized measurement.To solve the problem above, this thesis proposes an effective solution. The entire imaging measuring method is adopted, which is based on template. Image mosaicing and high precision image feature extraction are studied. And the selected methods are also implemented by software. The main research work in this thesis includes:1. The features of entire imaging measuring method are described. And different kinds of image mosaicing methods are compared. Finally, the mosaic method based on template is adopted. A template is designed, which is used for mosaicing four images. The algorithm for mosaicing four images is also deduced.2. Many kinds of edge detection methods are studied, considering the demand of the template based image mosaicing method and the precision of microparts measurement. The subpixel edge detection algorithms are mainly investigated. Experiments are performed to validate the algorithms. B spline curve fitting is selected to obtain the subpixel location of the edge.3. Image feature extraction algorithms are studied. The center-point and the radius of the circle can be obtained with high precision. Template matching, image segmentation, contour extraction and contour searching algorithms are implemented to get information from every single image before mosaicing.4. The algorithms for image mosaicing and measuring are implemented in Visual C++ environment. And man-computer interaction function for measuring is added in the software. Experiments for image mosaicing and measuring are carried out to validate the algorithms. The results indicate that the micropart entire imaging measuring method has achieved high precision.