A Review of Shadow Detection and Reconstruction in VHR Images

Abstract

Shadows in very high resolution (VHR) images represent serious problems for their full development. This paper faces this complete problem through the proposal of a processing chain, which is based on various advanced image processing and pattern recognition tools. We are using Canny Edge Detection Algorithm. Shadow is one of the most important problems in remotely sensed imagery which affects the accuracy of information extraction and change recognition. In these images, shadow is generally formed by different things, namely, cloud, urban and mountain resources. The shadow correction process consists of two steps: detection and de-shadowing. We are trying to review a range of techniques for both steps, focusing on urban regions (urban shadows), high areas (topographic shadow), cloud shadows and fused shadows. In recent years, thresholding and reconstruction techniques have become significant for shadow detection and de-shadowing, respectively. The purpose of edge detection is to considerably reduce the amount of data in an image, while protecting structural properties for further image processing. This worksheet focuses on a particular one developed by John F. Canny (JFC) in 1986 [14]. Even though it is quite old, it has become one of the standard edge detection techniques and it is still used in research.