STUDY ON SHADOW DETECTION FROM HIGH-RESOLUTION SATELLITE IMAGES USING COLOR MODEL

Abstract

Shadow detection is an important process for applications like classification, change detection, image interpretation, object detection, and recognition. The existence of shadow in satellite images reduce the amount of information that can be extracted and accordingly makes these applications more difficult or even impossible. Different color space is used to detect shadows based on , , and bands. This paper aims to represent an automatic approach for shadow detection from high-resolution satellite images. In this approach, a new index to highlight shadow areas based on color model is developed. A comparative study is carried out between the proposed index with a different photometric invariant color model, including IHS, HSV, YIQ, and models over the ratio and single-band images. Then, an automatic thresholding method is applied in images histogram. The accuracy of the obtained results is evaluated in terms of visual comparisons and shadow detection accuracy assessments. Experimental results show that the proposed ratio method provides the best results for shadow detection. On the other hand, shallow water is still misclassified as a shadow.