CLASSIFICATION ACCURACY ASSESSMENT EFFECTED BY SHADOW

Abstract

The biggest problem in Satellite Image of hilly area is the shadow due to terrain. A different methodology for supervised classification of satellite image is drafted. In this work we used Digital Elevation Model to model the terrain to extract the shadow information, which is a major cause for error in classification in mountainous area. The shadow depends on the terrain and the sun's position. The DEM can be used to model the terrain and once position of the sun is known, the area under shadow can be easily separated out. The irradiance mapping- (an extension to the shaded relief principle) is used. Sun Angle Calculation algorithm is implemented and incorporated in the irradiance mapping technique. The tool has been developed. for modeling and casting out the shadow area from the image. The image analyst now can easily get the shadow area from the image with automatic algorithm using DEM. This study finds some useful prospect for incorporation of DEM into Classification process. The work also brings the new approach of integrating GIS and Image analysis into one platform. The results got here provide promising results and scope for further work to be done in this field.