Proposed Algorithm for Shadow Identification and Classification in VHR Satellite Imagery

Abstract

The Very High Resolution (VHR) satellite imagery offers great possibilities for urban environments. Unfortunately, shadows cast by buildings in high-density urban environments obscure much of the information in the image, leading to potentially corrupted classification results or blunders in interpretation. The objective of this paper is to propose an algorithm for better identification and classification of shadows in VHR images. Two prime stages were carried out i.e. shadow identification and shadow classification. The shadow detection accuracy was calculated by using Cohen’s Kappa Coefficient and hence User’s Accuracy and Producer’s were also measured. It is observed that support vector machine gives better classification accuracy. The reconstruction of an original image is accomplished by using border interpolation. Experimental results are obtained on three VHR images representing different shadow conditions.

Keywords: Cast shadow, self shadow, shadow detection, VHR satellite images, support vector machine