An Algorithm for Boundary Extraction Based on Turning Row Permutation into Sequence Permutation  

作　　者：付梦印 张晓晨 周培德 

机构地区：[1]School of Automation,Beijing Institute of Technology

出　　处：《Journal of Beijing Institute of Technology》2010年第3期343-347,共5页北京理工大学学报（英文版）

基　　金：Sponsored by the National Natural Science Foundation of China (60773044 );PLA General Armament Department Key Foundation(9140A26010308BQ0178)

摘　　要：Image output by computer is done row by row,the shortcoming of which is that we cannot know the topological relations between points at the image boundary.To obtain the closed boundary of the target from an image and turn points＇ row permutation of boundary into sequence permutation,an algorithm based on 4-connection relation of pixel points is proposed and the analysis of its correctness and complexity is given.Compared with classical operators and existing method,the single pixel boundary extracted from gray images by this algorithm is more complete and more distinct.Moreover,it has a lower time complexity and can eliminate the noise inside and outside the target area.The obtained sequence permutation of points is also conducive to follow-up image processing.Image output by computer is done row by row,the shortcoming of which is that we cannot know the topological relations between points at the image boundary.To obtain the closed boundary of the target from an image and turn points＇ row permutation of boundary into sequence permutation,an algorithm based on 4-connection relation of pixel points is proposed and the analysis of its correctness and complexity is given.Compared with classical operators and existing method,the single pixel boundary extracted from gray images by this algorithm is more complete and more distinct.Moreover,it has a lower time complexity and can eliminate the noise inside and outside the target area.The obtained sequence permutation of points is also conducive to follow-up image processing.

关 键 词：4-connection sequence permutation boundary extraction infrared image 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]