Abstract:In order to solve the problem that the threshold is too high or too low in the traditional edge detection method, the key information is omitted or the interference information is mistaken as important information, resulting in unreliable edge detection results. By introducing the idea of adaptive threshold, this paper studies the edge detection method of trajectory image based on automatic parking. Histogram equalization, adaptive binarization and corrosion followed by expansion are used to eliminate small objects and smooth large objects. The gradient amplitude and gradient direction of different points in the pre-processed image are solved by first-order differential operator, and the ridge band in the gradient amplitude image is refined, only the local maximum of the amplitude is preserved. The logarithmic transformation method is used to expand the gradient range. A new local adaptive thresholding method is used to determine the thresholds and realize the automatic edge detection of parking track images. Hough transform is used to extract the straight line feature of the edge line segment of the trajectory image to obtain the straight line trajectory. The results show that the edge detection method is detailed and the overall performance is excellent.