Satellite Images Classification Using CNN :A Survey

Satellite images are taken in different conditions that are affected by the weather, the atmosphere, the noise they contain, and the difference in lighting. In addition, the images might be taken at night. For mapping complicated urban areas, detailed land cover information is useful. Recent advancements in satellite sensing technology promise data suitable for their intended use, especially when processed utilizing modern categorization techniques. Survey studies help create a knowledge foundation that helps researchers to find diverse facets of the main issue they intend to explore, mainly after being briefed on previous researchers' efforts, theoretical and methodological factors, and concepts and assumptions found in previous studies. In this regard, assumptions play a significant role in crystallizing the research topic without testing these hypotheses or proving their validity, which helps researchers develop and formulate the research topic using a considerably firm foundation. This paper reviews the search for a classification of satellite images using deep convolutional learning. Particularly, this research studies the different methods available in the literature, providing an overview of convolutional neural networks (CNNs) and some types of their models. Moreover, it provides a complete comparison and analysis of various research projects based on: CNN models trained on satellite image datasets to achieve accuracy. The most common CNNs used to classify satellite images are Alex Net, you only look once (YOLO), Visual Geometry Group VGG-16, U-Net, and residual learning Res Net. These networks have achieved different results for classification accuracy according to the used dataset.