Edge Detection of Noisy Images Based on Improved Canny and Morphology

Abstract

In the detection of image edge with noise, it is difficult for the traditional Canny algorithm to filter the noise, and its detection effect is poor. In order to solve this problem, the Canny operator and the morphological edge detection algorithm are improved in this paper. Then a fusion algorithm based on improved Canny operator and morphological edge detection is proposed. First of all, the algorithm improves the traditional Canny operator from three aspects: first, a hybrid filter consisting of an adaptive median filter and Gauss filter is designed; second, the gradient is calculated using the 4-direction Sobel operator; third, the global threshold segmentation algorithm is used to determine the threshold of edge connection. Then, the edge detection is carried out by the mathematical morphological method of multi-directional structure element and adaptive weight calculation. Finally, the resulting images from improved canny operators and improved morphological methods are combined by weighting to obtain the final edge image. Experiments are designed to compare noiseless images with Gaussian and salt and pepper mixed noises. The results show that the denoising effect of the fusion algorithm proposed in this paper is obviously better than that of the improved canny operator and the improved mathematical morphology method alone, and the edge graph is continuously clear and has good performance.