Rail Defect Edge Detection Based on Improved Canny Operator

Abstract

In view of the classical Canny operator's susceptibility to noise interference and the inadequacy of noise suppression and detection accuracy in rail image edge detection, an improved Canny operator based edge detection operator is proposed. The improved non-maximum suppression is used to further enhance noise suppression and edge pixel selection. The neighborhood mean Otsu algorithm is used to carry out high and low thresholds. Solve the problem and determine the high and low thresholds adaptively. The simulation results show that the improved algorithm can detect the edge details more perfectly in the rail image edge detection, and can also suppress noise, and has strong adaptability.