Research Of Image Edge Detection Based On Phase Congruency Algorithm

After the Fourier decomposition of the image,the phase congruency of the Fourier component of the signal is the largest on the part such as the step edge,the line segment and the Mach band.Compared with the image feature based on the luminance gradient,it has advantages over annotating the image features based on the phase congruency.Phase congruency is a non-dimensional value,which is not affected by the brightness and contrast of the image,so it can accurately measure the feature points of the image.This article mainly expounds how to calculate and extract image edge features by using phase congruency,and verifies the superiority of using phase congruency image feature extraction;and proposes a novel adaptive noise threshold phase congruency algorithm.It is proved by experiments that the applicability of PC algorithm and the accuracy of edge feature extraction are improved obviously without increasing tremendous time complexity.In this paper,the general framework is as follows:the first chapter is the summary of related research background,and the present PC related research results;the second chapter briefly introduces several kinds of methods based on traditional edge detection based on brightness gradient and their faced problems;the third chapter describes the theoretical basis of PC algorithm and its algorithm process steps;the fourth chapter puts forward the PC algorithm with adaptive noise threshold,and expounds its theoretical basis and the specific algorithm process.The fifth chapter applies the PC algorithm in the edge extraction of actual images of rotary coal kiln.Compared with the traditional characteristics of the brightness gradient detection,the article verifies the edge feature extraction algorithm for edge extraction is a better algorithm in small contrast and poor quality pictures.