Object Oriented Shadow Detection and an Enhanced Method for Shadow Removal

摘要

Remote sensing technology derives information about the earth resources from satellite imagery. In urban areas, mainly the elevated objects are the reason for the appearance of shadows. Shadows are created as the light source has been blocked by something. Shadows lower the quality of images or it may affect the information provided by them. Thus for the correct image interpretation or to extract information it is important to detect shadow regions and restore them. Significant researches has been going on in finding the best shadow detection and removal methods. Many algorithms and methods had been developed so far. Most of them are based on pixel level shadow detection and removal. An object oriented shadow detection and removal resolves the shortcomings of pixel level detection. Existing object oriented method suffers from color casting of images even after the correction. A new method for shadow compensation by image inpainting resolves these color casting problem.