基于PSO优选参数的SVR水质参数遥感反演模型

Model for water quality remote retrieve based on support vector regression with parameters optimized by particle swarm optimization algorithm

为进一步提高多光谱图像水质反演的精度,提出了一种基于PSO优选参数的SVR水质参数遥感反演模型。该模型利用高分辨率多光谱遥感SPOT-5数据和水质实地监测数据,采用CV估计模型推广误差,并使用PSO优选SVR模型参数,实现了模型参数的自动全局优选,在训练好的SVR模型基础之上对水质进行反演。以渭河陕西段为例进行实证研究,实验结果表明,所提出的水质反演模型较常规的线性回归模型有更高的反演精度,为内陆河流环境遥感监测提供了一种新方法。

In order to improve the accuracy of the water quality retrievals of multi-spectral image,the author puts forward a model for water quality remote retrieve based on support vector regression with parameters optimized by particle swarm optimization algorithm.The model uses high-resolution multi-spectral remote SPOT-5 data and the water quality field data,uses CV to estimate the promote error and use PSO to optimize parameters of SVR model.It optimizes the model parameters globally,after the water quality is retrieved by the trained SVR.The proposed model is applied to the water quality retrievals of Weihe River in Shaanxi province.The results show that the developed model has more accuracy than the routine linear regression model.The paper provides a new approach for remote sensing monitoring of environment to inland rivers.