Research On The Inspection Of Coordinate Of Tube’s Endpoints Based On Multi-vision

Pipelines are widely used in industrial areas such as aviation, aerospace, shipping and automotive, and the machining quality of pipelines directly affects the performance and reliability of a product. In order to guarantee the machining accuracy of a pipeline, it is necessary to accurately inspect the spatial shape of bent tubes and further adjust them. Meanwhile, the relative position accuracy of tube’s endpoints needs strict control to make sure pipelines can be assembled correctly. To solve the problem of accurate measurement of tube’s endpoints in pipeline system, an inspection method based on multi-vision is systematically researched in this paper, and a measuring system is developed. Several measuring instances are carried out to verify the performance of the proposed systemThe main achievements of the research are summarized as follows:1. The significance of measurement of tube’s endpoints is analyzed. The related fields at home and abroad are discussed. The problems of the current methods are summarized and the research purpose of this paper is pointed.2. By analyzing the relative theory of image processing in multi-vision-based tube inspection, a multi-vision-based inspection method for tube’s endpoints is proposed. The general process and key technique of this method is discussed.3. The illuminating and reflecting mechanism of multi-vision system is studied and a preprocess method of image based on reflection removing of tube’s surface is proposed. A photometric linearization method is adopted to remove specular reflection and diffuse region is located. Diffuse component is further removed by changing the intensity of diffuse pixel to the mean grey value of shadow region, which minimizes the effect of reflection to image processing and enhances the correctness of edge location.4. The property of tube’s centerline in an image is analyzed. An endpoint location algorithm based on centerline is proposed, locating the endpoint’s coordinate with the assistance of tube’s centerline. A conception which is called inclination of end-face’s projection in image is defined. This conception combines the centerline-based method with conventional contour fitting method together. The new method can improve the location accuracy of endpoints and is stable with the change of end-face’s pose in image.5. To make the research result into engineering application, multi-vision-based tube’s endpoints inspection system is developed. The developing background, technical way, function module and application method are expatiated. And the instance of the system is verified with an example.6. Finally, the content of the whole paper is summarized and the future works are discussed.