Modified Hue over Intensity Ratio Based Method for Shadow Detection and Removal in Arial Images

International audienceIn this paper, we represent an outline to automatically identify and remove shadows from the real single image. The most significant features can be manipulated by multiple convolution deep neural networks (ConvNets) process. The required features extraction methods are imposed at the dominant boundaries in each super-pixel level of the image. The shadow masks may be predicted random model to generate a smooth image based on Bayesian formulation. These works represents a chromaticity based process for removal of desire shadows in aerial images. Hue is considered as the base variables for shadow detection. We represent modified hue over intensity ratio method for the better result