LULC analysis of Bangladesh using Deep Learning

Land use land cover (LULC) is defined how the land is used by humans or what the surface is covered with. Specifically, land use means how a particular land is being used for; for example, farmland, built-up etc. Land Cover refers to what the surface is covered with; it can be river, forest, mountains etc. Thus, LULC information and LULC change over a period of time obtained from satellite data can be used to analyze the geography, socio-economic condition, poverty estimation, urban planning, building categorization etc. In this project we aim to obtain accurate LULC data from satellite images using deep learning methods. Deep learning based methods have been proved to be more effective to extract LULC information compared to traditional classifiers that are still used in GIS systems for this purpose. However, in order to train deep learning methods on satellite images we need to have a good amount of images annotated with ground-truth. For the LULC segmentation task we need pixel by pixel LULC class annotation for the satellite images. The goal of this project is to produce good quality LULC annotation data pixel by pixel for Dhaka city and surrounding area. With a sufficiently trained segmentation model, we can employ this model to perform LULC over the entire Bangladesh without any human effort. Moreover, this project targets to develop segmentation models that can outperform the existing ones by a significant margin.