水体叶绿素遥感反演中的多尺度分析

Multi-scale Analysis of Chlorophyll Inversion by Remote Sensing

水体遥感影像提取污染物信息普遍存在着尺度效应问题,选择合适的空间分辨率影像能够准确地表征水域污染的空间分布状况。充分利用水体空间尺度信息,研究水色遥感的尺度问题,有利于提升湖泊水体遥感反演模型的应用能力。以巢湖水域HJ-lA卫星HSI高光谱和CCD多光谱遥感数据为例,以Matlab为平台,采用基于离散小波多尺度变换分析方法,生成多光谱尺度和空间尺度影像,然后利用水体污染物的定量遥感反演方法,获取湖面水体叶绿素多尺度空间分布浓度,并利用水面同步实测数据对反演结果进行验证。通过结果比较,确定100 m分辨率的HSI高光谱数据为较优分析空间尺度。

Scale effect exists generally in water pollution information extraction from remote sensing image. The appropriate spatial resolution image can accurately characterize spatial distribution of water pollution. By making full use of the information of water spatial scale, solution on scale change of water color remote sensing is beneficial to improve the application ability of remote sensing inversion model of lake. Taking Chao- hu as an example, the paper applies HJ-1A HSI hyperspectral and CCD multi-spectral remote sensing data to build multi-spectral and spatial scales of image by discrete wavelet multiscale transform on the platform of Matlab, and then obtain the spatial distribution of chlorophyll concen- tration by the quantitative remote sensing inversion method of water pollutants, and finally verify the inversion results using water surface synchro- nous measurement data. By comparing the results, the 100 m resolution of HSI hyperspectral data is the better spatial scale of analysis.