Change detection method taking into account shadow information for high resolution remote sensing image

Abstract

Shadow is one of the interpretation keys to remote sensing images; nevertheless,shadow brings about disadvantageous effect on building change detection. The authors therefore deal with the problem of the removal of the error change detection results caused by shadow. The change detection method taking shadow information into account was proposed in this paper. First of all,the shadow was extracted from the image. Secondly,the shadow extracted was used to remove errors in initial change detection results. Finally,the better change detection results could be obtained. In the above process,the accuracy of shadow extraction is the key point. Through an analysis of shadow spectral information and geometry information,the authors made use of object oriented image classification to extract the shadow. The experimental results show that the method proposed in this paper can solve the problem of error detection caused by shadow and improve the accuracy of image change detection effectively.