Edge detection method based on gray-scale image of specific class

Abstract

The invention discloses an edge detection method based on a gray-scale image of a specific class. The method is improved on the basis of a traditional canny edge detection algorithm; therefore, sensitivity to Gaussian noise and salt and pepper noise can be effectively reduced, adaptivity to different images can be strengthened, and edge detection precision and accuracy can be improved. The method comprises the following steps that: Gaussian filtering and median filtering are employed to remove Gaussian noise and salt and pepper noise in an image; image enhancement is carried out on the filtered image; non-maxima suppression is carried out on the enhanced image as well as description and analyzing are carried out in a gradient histogram mode; adaptive setting on a high threshold value and a low threshold value of a detection and connection edge is realized; and an edge tracking algorithm is set to realize connection of edge points. According to the edge detection method provided in the invention, edge extraction under circumstances that Gaussian noise and salt and pepper noise interferences exist can be effectively improved; adaptivity to different images can be improved; and good edge detection precision and accuracy can be realized.

Description

technical field [0001] This patent relates to digital image processing technology, in particular to an edge detection method based on a specific type of grayscale image. Background technique [0002] Edge is the most basic feature of an image, and edge detection is of great significance to image analysis and understanding. Edge detection includes the following basic steps: filtering, enhancement, detection, and location. Traditional edge detection methods include Roberts operator, Sobel operator, Prewitt operator, Log operator and Canny operator, etc. With the development of mathematical theory and signal processing technology, new edge detection algorithms such as wavelet analysis-based edge detection methods, mathematical morphology-based edge detection methods, and fractal geometry-based methods have emerged in recent years. Among them, the canny edge detection algorithm based on optimization theory has the advantages of large signal-to-noise ratio and high detection ac...

AI Technical Summary

Problems solved by technology

The traditional canny operator is affected by various interference factors in actual processing, which will produce some false edges; at the same time, the traditional canny algorithm needs to set the high and low thresholds in the edge extraction process based on experience, and cannot The actual content automatically changes parameter values, lacks adaptability to different images, and it is difficult to eliminate the interference of local strong noise. While detecting false edges, some local edges with slow gray value changes will be lost

Therefore, there are still some limitations in specific applications

Images