Research on Water Body Extraction Method Based on GF-2 High Resolution Remote Sensing Image

For the characteristics of GF-2 satellite image data, the water information in Liujiaxia area is extracted by singleband threshold method, multi-band algorithm, normalized difference water index threshold method(NDWI) and the decision tree method combining singleband threshold method with shadow water index(SWI). At the same time, a decision tree classification method combining single-band threshold method and enhanced shadow water index(ESWI) is proposed.The comparison of the effects of these water extraction method is carried out. It is found that the decision tree classification method combining single band threshold method with ESWI can effectively eliminate the influence of most shadows, and the extraction effect is better than that of SWI. NDWI and multiband algorithm has a second extraction effect.The single-band threshold method has a poor extraction effect.