Research On Image Edge Detection Algorithm Based On Fractional Differential

Edge is usually represented as the sharp change of pixel gray value in digital image,which reflects the characteristics of image discontinuity and contains a lot of information.Therefore,image edge detection has become an important work before image segmentation and image fusion.For the shortcomings of existing edge detection algorithms,such as discontinuous edge extraction,poor anti noise performance,and unable to extract more effective texture details,this paper proposes an improved edge detection algorithm based on fractional differentiation combined with the theory of fractional differentiation.The final experimental results show that the algorithm not only effectively obtains the complete edge contour of the image,but also extracts the rich texture details in the image,and its effect is better than the traditional edge detection algorithm both subjectively and objectively.The main contents and results of this paper are as follows:(1)This paper introduces the basic theoretical knowledge of fractional calculus and digital image edge detection in detail,and briefly introduces the commonly used classical edge detection operators,new image edge detection algorithms and traditional fractional differential edge detection algorithms.This part lays a theoretical and experimental foundation for the following research.(2)For the shortcomings of classical detection algorithms,such as discontinuous edge extraction,poor anti noise performance and traditional fractional differential algorithm can not extract more texture details accurately and effectively,an edge detection algorithm based on fractional differential is proposed.From the effect of fractional differential on the signal,we can see that the attenuation of the low-frequency part is the weakest from zero to first-order differential,and the details of the lowfrequency part can be retained to the greatest extent by choosing the appropriate order;although the improvement of the high-frequency part by the first to second-order differential is less than that by the second-order differential,it also achieves the required improvement.Therefore,it is proposed that the high-frequency and low-frequency parts of an image can be detected by the above two order templates respectively,which can not only extract more complete edges,but also detect the edge texture details to the greatest extent,and the effect of edge extraction is better than other algorithms.(3)In order to avoid the loss of some edge details and noise amplification,the algorithm adds a parameter m at the center of the mask operator in the high-frequency part of the image;in order to extract more weak edges,retain more weak texture details and ensure the continuity of the edge,the algorithm adds a parameter n at the center of the mask operator in the low-frequency part of the image,and the edge extraction effect is better after adding parameters.(4)For the defect that the fractional differential mask operator needs to specify the order artificially,the algorithm uses information entropy and standard deviation as the image feature information of the order adaptive function in the high frequency part of the image,and uses information entropy and gradient mask value as the image feature information of the order adaptive function in the low frequency part of the image.Two functions are used to get the best order from the image feature information.The high and low frequency parts of the image are extracted respectively,and then the final edge is obtained by weighted fusion.The extracted edge details are rich.