Water Extraction Application Based On Object-oriented Method And Sentinel Data

Surface water resources is an important resource for people's survival and healthy and stable development of society.Water resources are relatively scarce and unevenly distributed,droughts and floods occur frequently in China.Effective extraction of surface water information is the basis of rational development of surface water body and timely disaster response.Use of high-resolution satellite imagery can achieve accurate,rapid and timely access to information on water resources in geographic information systems.This article takes Sentinel-IA SAR imagery and Sentinel-2A optical imagery as basic data to extract water information from different regions based on object-oriented methods.The main research is as follows:(1)Water extraction based on object-oriented method and Sentinel-2A data.Based on the object-oriented method and Sentinel-2A images to extract water body of Puding County,which is characterized by typical karst geomorphological features and fragile ecological environment,Meanwhile,take supervised classification,unsupervised classification and Normalized Difference Water Index to extract water body in the study area,were performed to compare the advantages of several extraction methods.The research shows that compared with traditional pixel-based extraction methods,the water extraction based on the object-oriented method can effectively remove the hillshadow,with a higher extraction accuracy,the overall classification accuracy Kappa coefficient of 89.46%.and compare the effect of water extraction with Landsat 8 and GF-1 image data sources,concluded that the water extraction based on Sentinel-2A data has higher precision.(2)Flooded monitoring application based on object-oriented method and Sentinel-IA SAR dataThis article takes Sentinel-1A SAR image as data source,based on object-oriented method established a flooded area extraction process,the flood dynamic change detection of different periods and flood submerge area extracted in the Huixian karst wetland in Guangxi was realized.compared with the traditional pixel-based water extracted method,the research show that Sentinel-1A SAR data has a great application potential in flood detection field.object-oriented method can effectively avoid the"salt and pepper phenomenon",have high precision in flooded area extracted and improved the utilization of spatial information.The results of this study indicate that:Sentinel data has advantages both of temporal and spatial resolution for a freely available source of remote sensing image data,and has a promising application in the field of water extraction.Water extraction based on object-oriented technology can fully exploit the rich spectrum,texture,and space information of remote sensing images.It can effectively solve the influence of mixed pixels of other features such as water and shadows on extraction accuracy,and avoid the "salt-pepper phenomenon" to a certain extent.It is also suitable for the extraction of medium-high-resolution Sentinel remote sensing data.Using Sentinel image as the data source,the extraction of water body information based on the object-oriented method has great application prospects.It can provide scientific basic data and technical support for regional water ecological civilization and sustainable development of water resources.