Research On Image Edge Extraction Based On Canny And Fuzzy Algorithm

With the development of computer science and related disciplines, image edge extraction technology has achieved vigorous development. Image edge refers to the image grey value of mutation, is one of the most basic characteristics of the image.It contains most of the image information, image edge extraction is an important technology in image processing problems. Is also one of the most fundamental problems in image processing. Their quality will directly affect the performance of the entire computer vision system, image edge of effective image features extraction for us are high-level description, recognition and understanding has an important supporting role. Image edge extraction in aviation, aerospace, satellite remote sensing images, medical diagnosis, images, and other fields has been widely used. Many scholars carried on the thorough research, do a lot of work, also has obtained certain research results. So far, have proposed many algorithms of image edge detection and extraction, but as a result of the actual image often contains the noise, different levels of noise and the edge in the spatial domain is characterized by the change of the grey value is bigger, is the high frequency component in frequency domain, this makes both is difficult to distinguish, brought difficulty to edge detection. At the same time due to the different algorithms for different types of images with different extraction effect, and the complexity of algorithm and the time is different, so the image edge detection is one of the classic technology to date has not been settled satisfactorily.Image edge detection method based on the comprehensive treatment, on the basis of the research status and development trend from the classic algorithms of image edge detection, first used in a variety of algorithm simulation results of different image edge extraction, analysis of the different performance of the algorithm. Secondly on the basis of the analysis combined with the advantages of different algorithms, the improved method is put forward. In this paper, main work is as follows:One is for classical Canny operator and adaptive edge positioning accuracy is bad, for different images, edge extraction of the performance of different faults, the Canny operator and edge gradient amplitude value in locating method was improved to realize the image edge extraction; In the process of improvement, in view of the Canny operator due to previous computation complex computation, the final time dosage larger defects, thus to improve the gradient amplitude value calculation method, in view of the classical Canny operator at the same time the lower threshold selection, low threshold is set too big to the edge of the smaller residual image edge mutations, in order to solve this contradiction, chose four threshold edge detection method.Second it is to classic fuzzy Pal. King algorithm in the process of image edge extraction is rigidly with grayscale of the image to be processed to determine the gray part of on any account, at the same time in the process of processing through the matrix transformation and fuzzy matrix is obtained by inverse transformation for the determination of maximum and minimum values of matrix element that would make a lot of original image edge information leak causing the image edge extraction is inaccurate; In order to improve these deficiencies, this paper will adopt the method of fuzzy threshold segmentation, according to the characteristics of the different regions of the image to get the image area threshold, and then USES the image edge enhancement, in turn, the image edge detection, that can avoid the loss of the image edge details. At the same time, the improved algorithm can effectively avoid the traditional fuzzy Pal. King algorithm of membership function.Third is based on the current evaluation standards for image edge extraction is scattered, often used in this paper, according to the literature of subjective evaluation and objective evaluation standard, to summarize these two standards, combined with the objective evaluation standards used in the quantitative analysis of mathematics, image edge extraction effect is given by calculation results of quantitative analysis, to achieve the unity of the subjective and objective. At the same time, the improved algorithm is proposed in this paper application of the evaluation system is analyzed, found that achieve results and theoretical analysis results are consistent.