Reduction of atmospheric and topographic effect on Landsat TM data for forest classification

被引:57
作者
Huang, H. [1 ,2 ,3 ]
Gong, P. [1 ,2 ,4 ]
Clinton, N. [4 ]
Hui, F. [1 ,2 ]
机构
[1] Chinese Acad Sci, State Key Lab Remote Sensing Sci, Inst Remote Sensing Applicat, Beijing 100101, Peoples R China
[2] Beijing Normal Univ, Beijing 100101, Peoples R China
[3] Chinese Acad Sci, Grad Sch, Beijing 100049, Peoples R China
[4] Univ Calif Berkeley, Div Ecosyst Sci, Berkeley, CA 94720 USA
关键词
D O I
10.1080/01431160802082148
中图分类号
TP7 [遥感技术];
学科分类号
081102 ; 0816 ; 081602 ; 083002 ; 1404 ;
摘要
The incident radiance in forested areas with rugged terrain varies greatly with the changes in solar elevation and azimuth, slope and aspect of the terrain, and the relative position of trees. The geotropic nature must be considered in the course of topographic correction. The Sun-Canopy-Sensor (SCS) model is introduced to substitute the cosine correction in a physical model. We used an atmospheric simulation code, MODTRAN, and a digital elevation model (DEM) to calculate the path radiance, downwards diffuse radiance and two-way transmittance of direct and diffuse light at different altitudes. Based on the atmospheric parameters derived above and the Lambertian assumption, surface reflectance in a forested area was retrieved from Landsat Thematic Mapper (TM) imagery using a revised physical model. Meanwhile, a smoothed DEM was used to assess the effect of noise on the DEM and misregistration between the DEM and the satellite imagery. Correlation analysis, spectral comparison between sunlit and shaded slopes and a support vector machine (SVM) classification were performed to assess the effect of the revised radiometric correction algorithm. Results indicate that the revised physical model with smoothed DEM is more adequate for forested terrain and more consistent spectra for similar vegetation under different illuminations can be obtained. Finally, higher classification accuracy of forested land can be achieved with the revised correction algorithm compared with the SCS correction and the original physical correction model.
引用
收藏
页码:5623 / 5642
页数:20
相关论文
共 41 条
[1]   COMPUTATION OF DIFFUSE SKY IRRADIANCE FROM MULTIDIRECTIONAL RADIANCE MEASUREMENTS [J].
AHMAD, SP ;
MIDDLETON, EM ;
DEERING, DW .
REMOTE SENSING OF ENVIRONMENT, 1987, 21 (02) :185-200
[2]  
[Anonymous], 1971, P 7 INT S REM SENS E
[3]   Remote sensing and geomorphometry for studying relief production in high mountains [J].
Bishop, MP ;
Schroder, JF ;
Colby, JD .
GEOMORPHOLOGY, 2003, 55 (1-4) :345-361
[4]  
Bishop R, 2002, WORLD CL PARASITES, V3, P23
[5]   The use of the Minnaert correction for land-cover classification in mountainous terrain [J].
Blesius, L ;
Weirich, F .
INTERNATIONAL JOURNAL OF REMOTE SENSING, 2005, 26 (17) :3831-3851
[6]   Revised Landsat-5 TM radiometric calibration procedures and postcalibration dynamic ranges [J].
Chander, G ;
Markham, B .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2003, 41 (11) :2674-2677
[7]  
CIVCO DL, 1989, PHOTOGRAMM ENG REM S, V55, P1303
[8]  
COLBY JD, 1991, PHOTOGRAMM ENG REM S, V57, P531
[9]   ATMOSPHERIC CORRECTIONS TO SATELLITE RADIOMETRIC DATA OVER RUGGED TERRAIN [J].
DOZIER, J ;
FREW, J .
REMOTE SENSING OF ENVIRONMENT, 1981, 11 (03) :191-205
[10]   SPECTRAL SIGNATURE OF ALPINE SNOW COVER FROM THE LANDSAT THEMATIC MAPPER [J].
DOZIER, J .
REMOTE SENSING OF ENVIRONMENT, 1989, 28 :9-&