广东飞来峡库区水深遥感

Water Depth Extraction from Remote Sensing Image in Feilaixia Reservoir

通过深入分析太阳光在水体中传输的物理过程,引入水体光学厚度概念,根据辐射传输理论,推导出水深遥感物理模型。以广东省飞来峡水库为例,采用SPOT5多光谱遥感数据,对其进行精确的辐射纠正、大气纠正,转化为符合地表实际反射率的反射率图像。经过水陆分离,选用合适的遥感波段数据,代入有关水体光学参数,应用水深遥感模型,计算出水域各像元水深值,通过图像处理软件绘制水深遥感制图。由于水面采砂、网箱养殖等人类活动的影响,水库采砂区水域水体浑浊以及库湾水域叶绿素浓度偏大,影响水深遥感精度,本文的模型综合考虑了水中泥沙和叶绿素对水深遥感的影响,并采用多波段数据求解水深,信息量丰富,精度得到大幅度提高。

Based on the physical mechanism of light transmission in water,the concept of water optical thickness is introduced into the transmission study,and then an integrated model for quantitatively calculating water depth by using remote sensing techniques was presented in this paper.Take Felaixia reservoir in Guangdong as a case study in this paper.Firstly,SPOT5 remote sensing data through radiation correction and atmospheric correction were converted into reflectance image in accordance with actual surface reflectance.Secondly,the information of waters was extracted by comparing the spectra of waters and lands,and then optical parameters of waters were substituted into the integrated models,selecting the appropriate bands of remote sensing,for calculating the water depths.Finally,the calculated water depths were input drawing software to generate the remote sensing map of water depths.The influence of sand mining and cage culture,such as turning muddy and high chlorophyll concentration of the reservoir waters,on the precision of water depth information has been considered in the model building in this paper,and hence multi-spectral remote sensing data were successfully applied to improve the calculation accuracy.