基于环境一号HSI高光谱数据提取叶绿素a浓度的混合光谱分解模型研究

Study on Spectral Unmxing Model of Chlorophyll-a Concentration Extraction Based on HJ-1 HSI Hyperspectral Data

随着遥感技术的应用推广以及对研究精度的要求提高,越来越多的研究注意到混合像元的问题。在水质遥感监测中传感器探测的水体辐射亮度值是纯水和各种水质参数辐射亮度值的叠加,混合像元问题严重影响了水质定量遥感反演的准确性。基于环境一号HSI高光谱数据,首先分析了混合光谱分解模型的物理基础,然后基于采样点浓度大小和PPI(纯净像元指数)方法在遥感影像上提取纯水和叶绿素a的端元波谱,并利用线性光谱分解方法得到叶绿素a的丰度值找丰度值与叶绿素a浓度值之间的统计关系,建立了叶绿素a浓度反演的混合光谱分解模型,且反演精度较高。本文为水质定量遥感提供了一种新的思路。

With the application of remote sensing technology and the improvement of the research accuracy, more and more attention is paid to the problem of mixed pixels.In water quality remote sensing monitoring, the radiation value of the water measured by the sensor is the superposition of the radiation value of the pure water and all kinds of water quality parameters.Accordingly, mixed pixel problems seriously affect the accuracy of the water quality retrieval.This study is based on HJ-1 HSI hyperspectral data.Firstly, the physical basis of spectral unmixing model is analyzed.Secondly, the endmember spectrum of pure water and chlorophyll-a is extracted based on the measured point concentration and PPI （pixel purity index）.Thirdly, chlorophyll-a abundance values are obtained by linear spectral unimixing method.Finally, find statistical relationship between the abundance values and chlorophyll-a concentration, and then spectral unmxing model of chlorophyll-a concentration extraction is established.The model has a high retrieval accuracy.This study provides a new idea for chlorophyll-a centration in quantitative inversion.