An edge detection algorithm for mine roadway image

In view of problem of bad edge extraction effect when traditional Canny algorithm processes fuzzy mine roadway image, the paper proposed an edge detection algorithm for mine roadway image based on wavelet transform and Canny algorithm. The algorithm firstly makes wavelet decomposition for original mine roadway image to obtain low-frequency image and high-frequency image, so as to avoid influence of fuzzy image on edge detection effect. Then it uses Canny algorithm to calculate first order differences of low-frequency image and high-frequency image to obtain gradient images of low-frequency image and high-frequency image and obtains edge images of low-frequency image and high-frequency image through calculating the maximum value of local gradient. When edge image of high-frequency image has discontinuity point, it detects eight-point neighborhood of the point in edge image of low-frequency image to find connection point to obtain complete edge detection image of mine roadway. The experiment result showed that the algorithm can detect more edge points of image with better edge connection effect comparing with traditional Canny algorithm.