Optimization application of Canny algorithm in gear image edge detection

Abstract

Aiming at the problem of false contour edges in gear defect detection by traditional Canny algorithm due to the influence of noise and illumination, an improved Canny algorithm for gear image edge detection is proposed. Firstly, the optimized guided filter algorithm is applied to the preprocessing of gear image, which improves the quality of image processing. Then the gradient values in eight directions are calculated, which makes the interpolation of non-maximum suppression more refined than the original algorithm. Finally, based on OTSU algorithm, the gray-gradient mapping function is constructed to determine the optimal threshold, which solves the limitation of the original algorithm to determine the threshold manually by experience. The experimental results show that the quality factor of the edge detection results of the proposed improved Canny algorithm reaches 0.868. Compared with the original Canny algorithm, the quality factor performance is improved by 13.51%, which proves the effectiveness of the improved measures proposed in this paper.