Study On Surface Water Extraction Method In Beijing-Tianjin-Hebei Region Based On Multi-source Remote Sensing Data

The per capita water resources in the Beijing-Tianj in-Hebei region are only 10%of the national average,and the problem of insufficient water resources for a long time has become a key bottleneck restricting the coordinated development of the Beijing-Tianj in-Hebei region.Therefore,it is necessary to accurately understand the dynamics of the spatial distribution of surface water in the Beijing-Tianjin-Hebei region in order to properly plan the allocation of water resources and guide the coordinated development of Beijing-Tianj in-Hebei.Satellite remote sensing has wide-coverage,high-frequency,multi-scale surface parameter perception capabilities,and has been widely used in surface water monitoring research.However,for the monitoring of surface water bodies in complex geographical environments in large areas,the existing methods of extracting surface water bodies based on satellite remote sensing have many deficiencies,such as poor applicability and low extraction accuracy.In order to overcome these deficiencies and obtain more accurate surface water body information,this paper proposes different surface water body extraction methods for different satellite remote sensing data such as Sentinel 1 and Sentinel 2,etc.,relying on the Google Earth Engine(GEE)data acquisition and processing platform,Conduct research on surface water extraction in the Beijing-Tianj in-Hebei region to provide support for surface water resources management in the Beijing-Tianj in-Hebei region.The main contents of this article are as follows:(1)A method for surface water extraction based on simple non-iterative clustering and random forest is proposed.This method uses simple non-iterative clustering image segmentation and random forest classification as the core,which can solve complex features in large-scale research areas.Under the conditions of surface water extraction,it can also achieve high-frequency,long-term sequence of surface water monitoring and extraction;(2)The optical image water extraction method based on the HSV color space is improved,and a surface water extraction algorithm for the Beijing-Tianj in-Hebei region suitable for the Sentinel 2 optical image is proposed.The color space is the core,constructing characteristic equations and extracting surface water bodies.It has been verified that the overall extraction accuracy is higher than 90%,which can achieve high-precision surface water extraction;(3)Using the extraction method based on simple non-iterative clustering and random forests,the surface water bodies in the Beijing-Tianjin-Hebei region are monitored and extracted once a month for the submerged area from July 2015 to December 2019.At the same time,the average monthly inundated area in the year is calculated according to the year,and the scale of the surface water in the current year is represented by the monthly average inundated area.The change trend obtained is basically consistent with the data in the Water Resources Bulletin;(4)Using the extraction method based on the HSV color space,the annual flooding frequency extraction analysis of the surface water bodies in the Beijing-Tianj in-Hebei region was conducted from 2016 to 2019,and the frequency of occurrence of the annual water bodies was calculated.The results show that the total area of short-term submerged water in the Beijing-Tianjin-Hebei region shows a decreasing trend.The total area of submerged water in seasons remains basically stable,and the total area of long-term submerged water increases year by year.