Shadow Identification From The Dense Pixeled Satellite Images

Information from the high-resolution satellite images can be extracted much more efficiently of present shadows of the objects or clouds from those images can be detected and removed. These shadow regions cause loss of information and sometimes get mixed with the darker objects. n satellite image processing, shadow occurrence may cause to deteriorate the information in the image. When there is any obstacle in the path of light, it may cause to shadow generation. In satellite images it occurs more as the position of the sun-rays keeps change their angle of elevation, and thus shadow created at different angles at different point of the time. Cloud shadows may also cause uncertainties in applications like Landmapping, surveillance, weather prediction, etc. In the proposed methodology, a dataset is created including some High-resolution and some Low-resolution satellite images, and using their different channels, like HSV, and extraction of their last available band, some shadow masks are generated.