Machine Learning Techniques Applied of Land Use—Land Cover (LULC) Image Classification: Research Avenues Challenges with Issues

Abstract

An easy-to-use programming environment, open access to satellite data, and access to high-end consumer computation power has made it very easy to align remote sensing and machine learning during the new era. A variety of remote sensing applications have utilized publicly available data. Land use (LU) image classification has become vitally important in the natural environment because of the expansion of some global changes relating to the temperament of the earth. Therefore, researchers should investigate this area more deeply. This paper presents a complete review to help out the researchers to carry on with the land use/land cover for image classification process, as there are limited numbers of review articles to assist them. The purpose of this paper is to discuss the classification of satellite images using mainly employed machine learning algorithms. We discuss the general process of LU/LC based on multi-image classification, as well as the challenges and issues faced by researchers. Only a few studies evaluate machine learning algorithms for image classification using openly available data, however.