Deep Learning Techniques for Crop Classification Applied to SAR Imagery: A Survey

Crop classification is a major concern to the growth of a country. Deep learning (DL) is one of the significant decision supporting tools for cataloguing and decision making on the type of crops at a particular time or season. Various machine learning (ML) and DL techniques are implemented for crop classification. Due to cloud free, all weather and light condition working characteristics of SAR., it has gained tremendous potential in crop classification. In this review paper we excerpt some of the algorithms and geographies that have been implemented in classification of crop for SAR remote sensing. Here we have reviewed almost 50 relevant studies from various database., amongst these some are selected for future studies. We initially discuss the use of classification of crop and the imaging techniques., followed by the development in technology by applying contemporary techniques to get better results. We also discuss on the comparative performance analysis on the techniques implemented earlier that helps us to frame the research questions and identify the gaps for further research in this area. Convolution neural network (CNN) is the utmost used DL algorithm for crop classification.