Water Extraction Based on Landsat Remote Sensing Image

Abstract

With the rapid development of remote sensing information technology, the types of Earth observation products are increasingly diverse, and the spatial and temporal resolution of remote sensing images are greatly improved. In recent years, how to extract useful information from massive remote sensing data products is a hot issue in remote sensing geoscience research, among which the extraction of water information can be widely used in agricultural production, water resources protection and monitoring, disaster prevention and reduction and other applications. Based on the characteristics of remote sensing information extraction, this paper analyzes the water information extraction of remote sensing images from shallow to deep. Threshold method and normalized water index method are respectively used for water extraction, and comparative analysis is conducted. The results show that both threshold method and normalized water index method can effectively extract water, but the threshold method requires several experiments to determine the threshold value. It is impossible to determine the most suitable threshold for water extraction. Normalized water index can extract water well, and it is realized by the operation between bands, and the operation is simple and convenient.