基于高分一号卫星数据的库区淹没频率分析方法

为分析水库水位变化对消落带生态结构的影响,提高淹没频率计算精度,提出一种基于高分一号卫星数据的淹频率分析方法。该方法通过对高分一号WFV影像数据进行预处理并多次迭代提取缓冲区的水体信息,结合地理配准和计算得到淹没频率空间分布特征,通过以此模拟的水体分布结果的精度验证淹没频率的可靠性。三河口水库实例验证结果表明:提出的水库淹没频率分析方法较准确地反映了水体信息的时空变化特征,模拟的水体信息结果总体精度和Kappa系数均值分别达到86.5%和73%,较准确地还原了水库蓄水阶段的水体变化,提高了水体淹没频率计算的精度。

An analysis method of reservoir inundation frequency based on GF-1 satellite data

In order to analyze the impacts of water level changes on the ecological structure of water-level-fluctuating zone in the reservoir and improve the accuracy of the calculation of the inundation frequency, an analysis method of inundation frequency is proposed based on GF-1 satellite data. Firstly, the GF-1 WFV satellite imageries are preprocessed and a buffer zone is set up to extract the water body through multiple iterations. Then, the spatial distribution characteristics of the inundation frequency are obtained by geographical registration and inundation frequency calculated from the proposed method. Finally, the reliability of the inundation frequency is verified by the accuracy of water body distribution simulated from the method. The application in Sanhekou Reservoir show that the proposed analysis method of the inundation frequency can accurately reflect the spatiotemporal variation characteristics of the water body information. The overall classification accuracy and Kappa coefficient of simulated water body results reach 86.5% and 73% respectively, which can accurately reproduce the water body changes during the impoundment period of reservoir and improve the accuracy of inundation frequency calculation.