Automated Shadow Compensation in Color Aerial Images

In urban aerial images, shadows cast by culture features due to solar orientation are usually homogenous and among the darkest regions in images and they make the feature under them unrevealed. This paper presents automatic procedures for shadow detection and radiometric restoration of features under cast shadows in color aerial images for solving problems caused by cast shadows in digital image mapping. This paper investigates the assumptions and conditions of shadow occurrence in urban color aerial images and employs color transformation, spectral ratioing, automatic thresholding, mathematical morphology, and local normalized histogram matching processes under physical model without the requirement of any a priori geometric knowledge or information about the scene and the source of illumination. Experimental results show that the proposed approach not only distinguish shadow regions with high accuracies but also compensate the luminance with color constancy and dramatically improved visualization of features within shadow regions in color aerial images.