论文部分内容阅读
高光谱遥感应用于内陆湖泊水质监测具有较好的发展前景,但由于内陆湖泊水体光学环境的时空多变性,如何高效利用水体高光谱特征信息,降低数据冗余度,发展高精度的水质参数反演模型具有重要的意义。针对上述问题,以巢湖为例,将遗传算法和地统计学相结合,利用环境一号(HJ-1A)卫星HSI高光谱遥感数据,建立了基于协同克里格遗传算法的湖泊水质总磷浓度高光谱遥感反演模型。实验结果显示,与传统遗传算法比较,协同克里格遗传算法模拟的ME、RMSE分别提高了128.2%、53%。经总磷实测值和反演值比对,建模和检验的相关系数R2分别为0.85、0.77。反演结果表明:协同克里格遗传算法通过利用克里格插值对传统遗传算法目标函数优化改进,使其具备克里格最佳局部估计能力,能够有效提高反演的精度。
Hyperspectral remote sensing has good prospects for inland lake water quality monitoring. However, due to the spatiotemporal variability of the optical environment in inland lakes, how to efficiently use hyperspectral features, reduce data redundancy and develop high-precision water quality Parameter inversion model is of great significance. In view of the above problems, taking Chaohu Lake as an example, the genetic algorithm and geostatistics are combined. Based on the HSI hyperspectral remote sensing data of HJ-1A satellite, the total phosphorus concentration of lake water based on the collaborative Krieger genetic algorithm Hyperspectral remote sensing inversion model. Experimental results show that compared with the traditional genetic algorithm, the ME and RMSE of the cooperative kriging genetic algorithm are increased by 128.2% and 53% respectively. The correlation coefficients R2 of modeling and testing were 0.85 and 0.77, respectively, after comparing the measured value of total phosphorus and the inversion value. The inversion results show that the cooperative kriging genetic algorithm can improve the objective function of traditional genetic algorithm by using kriging interpolation and make it have the best local kriging ability, which can effectively improve the accuracy of inversion.