Support Vector Machine based Remote Sensing using Satellite Data Image

Shadow makes the Image and Video critical and hence it is very mandatory step to remove the shadows especially in terms of Satellite Images. The method detects and subtract shadows from a Satellite Image. Our method’s objective is to detect and remove Shadow from Satellite Images by learning the features of the super pixels (pixels with same characteristics) and the edges of the image using Supervised Machine Learning model namely Support Vector Machines(SVM).Using a statistical model like CRF model to create the mask of the shadow. The features of shadow are extracted from the Images followed by a SVM classifier to precisely collect the shadow matte followed by the collection of shadows. This Shadow removal method is performed using Bayesian formulation and evaluated by Gray Level Co-Occurrence Matrix.