Updating of GIS maps with Change Detection of Buildings using Deep Learning techniques

In this rapidly developing world, most of the significant growth can be seen in the urban settlements. To keep a record of the extent of this growth, change detection of buildings in the urban areas can be done. With the advent of technology, access to very high resolution satellite images became easier and two temporal images of the same location can be used to detect changes. GIS mapping transforms geographical data into digital maps that help in easily identifying patterns, trends and relationships. In the proposed methodology, deep learning model such as STANet will be used in which features will be extracted from the bi-temporal images. Residual Neural Network (ResNet) acts as a backbone to this model. The first module will focus on the change detection, calculation of built-up areas changed between the bi-temporal images. The second module will update GIS maps with the changes detected.