| The edge is the most basic information feature of the image,and also the region with t he most abundant image information,which represents the intensity discontinuity in the im age.Edge plays a crucial role in image processing.On the one hand,it can well retain the s hape information of the object in the image,on the other hand,it can reduce the original inf ormation of the image,thus reducing the information to be processed.Therefore,edge dete ction of the image is one of the most necessary research fields.The information content of2 D images can be well reduced by edge detection processing.In most high machine vision fields,the edges of images can be further analyzed,such as motion analysis and target anal ysis.Firstly,edge detection,several mature edge detection methods and their advantages a nd disadvantages are described.This paper gives some theoretical knowledge based on Ga bor function again,focusing on Gabor transform,Gabor filter,wavelet transform and Ga bor wavelet transform,so that readers can easily distinguish.Gabor filter is able to achieve joint minimum uncertainty,which was proved by J.G.Daugman.Based on these results,th e researchers proposed a computational model based on the Gabor filter and applied it to th e underlying visual system,such as edge detection,texture classification and data compress ion.The edge detection evaluation proposed by Canny was used to study the edge detectio n performance based on Gabor odd filter.On this basis,the design criterion of Gabor odd fi lter edge detection is deduced.The results show that the simulation results based on its crit eria prove the properties and effectiveness of edge detection based on Gabor odd-part filter model.Previous researches on edge detection of Gabor complex filters have been conduct ed by researchers,but few odd parts of Gabor complex filters have been singled out for edg e detection,in order to improve the operation speed under the premise that the detection eff ect is almost the same. |
PDF Full Text Request