Image edge detection based on beamlet transform  

作　　者：Li Jing Huang Peikang Wang Xiaohu Pan Xudong 

机构地区：[1]China Aerospace Science 8z Industry Corporation, Beijing 100854, P. R. China

出　　处：《Journal of Systems Engineering and Electronics》2009年第1期1-5,共5页系统工程与电子技术（英文版）

摘　　要：Combining beamlet transform with steerable filters, a new edge detection method based on line gradient is proposed. Compared with operators based on point local properties, the edge-detection results with this method achieve higher SNR and position accuracy, and are quite helpful for image registration, object identification, etc. Some edge-detection experiments on optical and SAR images that demonstrate the significant improvement over classical edge operators axe also presented. Moreover, the template matching result based on edge information of optical reference image and SAR image also proves the validity of this method.Combining beamlet transform with steerable filters, a new edge detection method based on line gradient is proposed. Compared with operators based on point local properties, the edge-detection results with this method achieve higher SNR and position accuracy, and are quite helpful for image registration, object identification, etc. Some edge-detection experiments on optical and SAR images that demonstrate the significant improvement over classical edge operators axe also presented. Moreover, the template matching result based on edge information of optical reference image and SAR image also proves the validity of this method.

关 键 词：edge detection beamlet transform steerable filters optical image SAR image. 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TU753[自动化与计算机技术—计算机科学与技术]