This article aims to proposea circular edge detection algorithm based on visual image processing for the accurate counting of straight pipes, so as to improve the counting accuracy and packaging efficiency.CCD was used to collect the end image of straight pipe, and the detection algorithm was written in Matlab software. The contrast of the image was enhanced by adjusting the coordinates of the pixels and the size of the threshold, and the gray and binary preprocessing of the image werecompleted. Hough transform and Harris Laplace algorithm werecombined to detect the edge of the end surface of the pipe. The circular edge of the end surface was recognized, displayed and counted by setting the diameter of the pipe. The straight pipes wereput into the designed visual counting device. After the end section was aligned by the movable baffle, the visual counting was realized by image processing technology. The correctness of the system was verified by many test results. This method can be used for automatic counting of straight pipe sections.