In order to solve the route deviation problem caused by industrial automated guided vehicle(AGV) navigation with cross ribbon as the characteristic sign, a positioning correction system based on XLD matching method was proposed by using Halcon machine vision software.The template image was preprocessed to eliminate shadows and impurities. A new mathematical model was created by using Gaussian derivative, which was filtered to obtain a clearer XLD profile and a more accurate coordinate of the center point of the cross ribbon, and then the XLD profile template was created; the XLD contour template matching was carried out to obtain the position information of the cross ribbon of the detected image and the positioning and correction were realized. The results show that by using the XLD contour shape matching method, the position information of the cross ribbon can be quickly, accurately and stably extractd, the average accuracy of the single detection image is 96.93%, the average matching time is 29.88 ms, and the correction is accurate, the method has high robustness.