PROJECT TITLE :

Shadow Detection of Man Made Buildings in High Resolution Panchromatic Satellite Images - 2014

ABSTRACT:

High-resolution satellite imagery is considered an excellent candidate for extracting information regarding the human activities on Earth. The information concerning residential development and suburban area mapping is of interest that may be obtained from these pictures. Shadow of structures like man-made buildings is one amongst the most cues for structure detection in panchromatic high-resolution satellite imagery. However, to correctly exploit the data of the shadow in a picture, the shadow needs to be detected and isolated first. During this paper, we propose a replacement algorithm for shadow detection and isolation of buildings in high-resolution panchromatic satellite imagery. The proposed algorithm relies on tailoring the traditional model of the geometric active contours such that the new model of the contours is systematically biased toward segmenting the shadow and the dark regions within the image. The systematic biasing within the proposed contour model is accomplished by novel encoding of the radiometric characteristics of the shadows regions. After detecting and segmenting the shadow and therefore the dark regions in the image, additional processing steps are introduced. The proposed postprocessing is predicated on choice of optimal threshold and a boundary complexity metric to distinguish the true shadows from the litter. Experimental results are presented to validate the performance of the proposed algorithm on real high-resolution panchromatic satellite images.