Machine Learning Algorithms for Classifying Land Use and Land Cover

We are in the Big Data era, the quantity of geospatial data gathered or stored using remotely sensed satellite imagery for land-use and land-cover (LULC) mapping and secondary geospatial data files grow. Innovative cloud computing, deep learning methods, and machine learning have also recently been developed. Deep learning algorithms got a lot of interest because of their enhanced performance in separation, category, and supplementary machine algorithm applications. Land use and land cover (LULC) are key elements of a broad range of ecological uses in remote sensing. Land-use changes occur on the geographical and spatial scale owing to the precision, development capabilities, flexibility, uncertainties, structure, and ability to incorporate existing patterns. As a result, the high performance of LULC modelling necessitates the use of a broad range of pattern modes in remote sensing, including dynamical, statistic, and deep learning models. Advances in remote sensing technology, and hence the rapidly increasing amount of timely data accessible on a worldwide scale, open new possibilities for a variety of applications. This article provides a summary of the basic ML and DL ideas that apply to the LULC is explained, including their pros and cons. To address the difficult issue of identifying changes in LULC, the application of deep learning to land usage, and this review has clarified both their advantages and disadvantages as researched by various scholars.