Classification Techniques in Remote Sensing: A Review

In order to find building supplies and offer precise geographic data, remote sensing is mostly utilized to study potential dam, bridge, and pipeline locations. Images taken by satellites and drones are used in image remote sensing analysis to study the Earth’s surface. Any classification method’s primary goal is to give semantic labels to photos that have been collected. Using these labels, the images may then be sorted in a semantic order. In many areas of remote-sensing, image retrieval and video analysis, the semantic layout of images is used. Early approaches to remote sensing picture analysis were built on the extraction and representation of low- and mid-level features. By utilizing feature optimization and machine learning algorithms, these systems have demonstrated good performance. Small-scale image datasets were utilized in these previous methods. Deep learning models are now being used more frequently for remote sensing picture analysis. The employment of multiple hybrid deep learning algorithms has demonstrated significantly better outcomes than the previous models. A thorough analysis of historical patterns is provided in this review paper, utilizing conventional machine learning principles. For the purpose of remote sensing visual analysis, a list of publicly accessible image benchmarks is also provided.