An Image Edge Detection Algorithm Based on Improved Canny
Kewen Liu, Xiao Kang, Hongxia Xiong · 2017
In view of the traditional Canny edge detection algorithm in the smoothing filter will make the image become blurred and threshold selection is not adaptive, and its defects in gradient calculation, this paper proposes an improved Canny edge detection algorithm.The algorithm uses adaptive median filtering instead of Gaussian filtering.The gradient method is adopted to enhance the influence of the middle pixel to improve the edge detection precision.The Otsu algorithm is introduced to calculate the threshold value according to the image pixel information, which avoids the difficulty of artificially setting the high and low thresholds.The simulation results show that the improved Canny algorithm has good anti-noise performance, and it has stronger adaptability when detecting more edge details. Traditional Canny Algorithm Principle of Algorithm.The Canny operator is implemented in four parts: image smoothing, calculating the gradient amplitude and direction, Non-maxima suppression, double-threshold detection and connection of edges.The basic flow diagram is shown in Fig. 1.