辛鑫等：Automated detection of new impact sites on Martian surface from HiRISE images

Automated detection of new impact sites on Martian surface from HiRISE images

作者:Xin, X (Xin, Xin)[ 1,2 ] ; Di, KC (Di, Kaichang)[ 1 ] ; Wang, YX (Wang, Yexin)[ 1 ] ; Wan, WH (Wan, Wenhui)[ 1 ] ; Yue, ZY (Yue, Zongyu)[ 1 ]

ADVANCES IN SPACE RESEARCH

卷: 60  期: 7  页: 1557-1569

DOI: 10.1016/j.asr.2017.06.044

出版年: OCT 1 2017

摘要

In this study, an automated method for Martian new impact site detection from single images is presented. It first extracts dark areas in full high resolution image, then detects new impact craters within dark areas using, a cascade classifier which combines local binary pattern features and Haar-like features trained by an AdaBoost machine learning algorithm. Experimental results using 100 HiRISE images show that the overall detection rate of proposed method is 84.5%, with a true positive rate of 86.9%. The detection rate and true positive rate in the fiat regions are 93.0% and 91.5%, respectively. (C) 2017 Published by Elsevier Ltd on behalf of COSPAR.

通讯作者地址: Wan, WH (通讯作者)

Chinese Acad Sci, Inst Remote Sensing & Digital Earth, State Key Lab Remote Sensing Sci, 20A,Datun Rd, Beijing 100101, Peoples R China.

地址:

[ 1 ] Chinese Acad Sci, Inst Remote Sensing & Digital Earth, State Key Lab Remote Sensing Sci, 20A,Datun Rd, Beijing 100101, Peoples R China

[ 2 ] Univ Chinese Acad Sci, Beijing 100049, Peoples R China