基于LOG算子的自适应图像边缘检测方法_严国萍.pdfVIP

DOI :10.13245 /j .hust.2008.03.02 36　3 　　　　　　　 () Vol.36 No.3 2008　3　 J.Huazhong Univ.of Sci.Tech.(Natural Science Edition) 　M ar.　2008 LOG 严国萍　戴若愚　潘　晴　刘袁缘 (华中科技大学 电子与信息工程系, 湖北武汉430074) :-(LOG), , LOG , , , , , LOG .. 　　:;;-; :TN9 .73　　:A　　:67 -45 2(2008)03-0085-03 Adaptive approach to image edge detection by Laplacian of Gaussian operator Yan Guop ing 　Dai Ruoy u 　Pan Qing 　L iu Yuany uan (Department of Electronic and Information Engineering, Huazhong University of Science and Technology, Wuhan 430074, China) Abstract:Because of its isotropy, traditional Laplacian of Gaussian (LOG)operator is not used in many occasions.The LOG operator w ith anisotropy is presented, w hich is sensitive to multi-direction and can intensify the edges that has same direction as the major axis.An adaptive algorithm w as de- signed to simulate the process of the human visiual system (HVS)concentration, and the results be- tw een the isotropy and anisotropy LOG operator are presented.The algorithm not only gives a ration- al and better explanation of the HVS , but also makes the traditional edge detection theory w ith LOG operator more perfect. Key words:image processing;computer vision;Laplacian of Gaussian operator;edge detection 　　-(LOG) , 1　LOG ., , , , G(x, y, σ), LOG . 2 G(x, y, σ), G(x, y, LOG , σ), σ, : [ , 2] 2 ., LOG G(x, y , σ)= (/(2πσ))exp (-(/2)A ); .