ARS | Publication request: RECOGNITION OF SEMI-ARID VEGETATION TYPES BASED ON MISR MULTI-ANGULAR OBSERVATIONS AND SURFACE ANISOTROPY PATTERNS INVERSED BY BIDIRECTIONAL REFLECTANCE MODELS

Submitted to: International Society for Photogrammetry and Remote Sensing Proc
Publication Type: Proceedings
Publication Acceptance Date: October 17, 2005
Publication Date: October 17, 2005
Citation: Su, L., Chopping, M.J., Rango, A., Martonchik, J.V., Peters, D.P.C. 2005. Recognition of semi-arid vegetation types based on MISR multi-angular observations and surface anisotropy patterns inversed by bidirectional reflectance models. In: Proceedings of the 9th International Symposium on Physical Measurements and Synthesis in Remote Sensing (ISPMSRS 2005), October 17-19 2005, Beijing, China. p. 186-189.

Technical Abstract: Mapping accurately community type is one of main challenges for monitoring semi-arid grasslands with remote sensing. Multi-angle approach has been proved useful for mapping vegetation types in desert. Multi-angle Imaging Spectro-Radiometer (MISR) provides 4 spectral bands and 9 angular observations. In this paper, several classification experiments were done to find the optimal combination of MISR multi-angular observations to mine the information carried by MISR data as effective as possible. The experiments show the following findings: 1)The combination of MISR 4 spectral bands nadir observation and red and near infrared bands C, B, A camera observations can obtain the best vegetation type differentiation at community level in New Mexico desert. 2)The k parameter at red band of Martonchik-Rahman-Pinty-Verstraete (MRPV) model and the structural scattering index (SSI) can bring additional useful information to land cover classification. 3)The information carried by the two parameters, however, is less than that carried by surface anisotropy patterns described by the MRPV model and a linear semi-empirical kernel-driven bidirectional reflectance distribution function model, RossThick-LiSparse-Reciprocal model. These experiments prove that multi-angular data raises the classification accuracy from 45.4% of nadir observation to 60.9%, and with surface anisotropy patterns derived from MRPV and RossThick-LiSparse-Reciprocal accuracy 67.5% can be obtained when maximum likelihood algorithms are used. Support vector machine algorithms can raise the classification accuracy to 76.7%. This research suggests that multi-angular observations, surface anisotropy patterns and SVM algorithms can improve semi-arid vegetation type differentiation remarkably.