Bag of Visual Words Methodology in Remote Sensing—A Review

With the increasing popularity and use of geographical data processing, more and more research efforts have been placed on geographical scene classification. Not only in the field of satellite imagery, but considering the recent overall, the overwhelming increase in the amount of visual information such as digital images present, there’s an urgent need of developing a model that aids in complete representation of an image and its features and also its retrieval. To obtain a complete representation of satellite images from a high spatial resolution satellite, bag of visual words (BoVW) model is emerging as a significant tool with quite promising results in the fields of image classification and image retrievals. This paper gives a detailed review of the introduction of BoVW methodology in the field of satellite imagery, its implementation, results, and future scope of BoVW method in remote sensing science and technology.