Accurate Detection of Buildings from Satellite Images using CNN

Satellite is a physical object that revolves around a larger object (celestial bodies) in space. The satellite are launched and orbited for a variety of purposes such as weather forecasting, GPS, communication, etc. Images obtained from satellites play a significant role in the field of remote sensing. The remote sensing image classification have found various uses in land monitoring, terrain feature classification, urban planning, groundwater exploration, environmental disaster assessment, etc. This paper describes the development of an accurate building detection system from the satellite images. In the proposed work, various methods are developed to detect the buildings in satellite images. It is found that the proposed method successfully identify various types of building footprints. The necessary algorithm for implementation of building detection is based on deep learning using the convolutional neural network model, which is constructed efficiently for the training and validation of the input dataset. The efficiency of the designed method is tested by repeating the experiment multiple times successfully. The proposed method produces a building detection with better accuracy of 83%.