Water Body Extraction of Remote-Sensed Image Based on D-S Evidence Theory
This paper proposes a method of water body extraction by the theory of D-S evidence. Firstly, for Thematic Mapper (TM) remote-sensed image, Mahalanobis distance, Bhattacharyya distance, and the ratio of band 5 to band 2 were used to obtain three different sources of evidence. Secondly, BPAF (Basic Probability Assignment Function) was obtained by different evidential sources respectively. Finally, extraction of water body in TM remote-sensed image was obtained by D-S evidence combining rules and target determining object principle. Dessau city Data was used to test our proposed method, and to compare with the water-body-extract methods respectively, which proves the proposed method more accurately. The developed method has two features: (1) Low fault tolerance; (2)High reliability as different evidential source data is fused.
Chenrong Huang Fengchen Huang Aiye Shi
School of Computer Engineering Nanjing Institute of Technology Nanjing, China, 211167 College of Computer and Information Engineering Hohai University Nanjing, 210098 P.R.China, College of Computer and Information Engineering Hohai University Nanjing,210098 P.R.China
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）