| This paper is a practical research project of oil-pipes’counting in LiaoHe oil-field. This paper has proposed technical methods on counting of oil-pipes based on Image Processing. The thesis has done much research on digital image processing, splice, and oil-pipes count recognition. And reasonable solutions to insure the accurate counting of oil-pipes have been found finally.In the image processing parts, this paper mainly researches the image enhancement and segmentation. The enhancement effect in the images of tubing has been compared and analyzed. The paper also has researched the edge detection algorithms, such as Roberts operator, Prewitt operator, Sobel operator, Krisch operator and LOG operator, and made a comparative analysis on them.In the part of image splice, this paper has used two set of line scanning cameras to get the tubing end images, and focused on the Phase correlation algorithm and the SIFT algorithm. Finally a combination of phase correlation SIFT algorithm which matched the images in the overlapping area has been used, and the method of coming into and get out gradually has been used to join the images together. In addition, the simulation experiments have been made by Matlab to analyze the accuracy and efficiency of the algorithm.In the round recognition and counting parts, this paper has used the improved randomized Hough transform to obtain the circles’center coordinates and round radius. Finally the method of reference has been used to complete the classification recognition. In the part of tubing counting, this paper has added OpenCV to provide image detection, image processing library, which could realize very rich functionality. And it has realized image processing and computer vision in many common algorithms. Finally it has used VC++to achieve the count of tubing. Furthermore, in this paper, some experiments have been done about the system to show that it can achieve the count of tubing in different situations. |
PDF Full Text Request