基于最小二乘支持向量机回归的背景偏振光谱二向反射分布建模分析  

Spectropolarimetric Bidirectional Reflectance Distribution Modeling for Background Based on LS-SVM Regression

作　　者：赵永强[1] 柳丹[1] 陈超[1] 程咏梅[1] 

机构地区：[1]西北工业大学自动化学院,西安710072

出　　处：《光子学报》2011年第10期1494-1499,共6页Acta Photonica Sinica

基　　金：国家自然科学基金(No.61071172;No.60602056;No.60634030);航空科学基金(No.20105153022);西北工业大学基础研究基金(No.JC200941)资助

摘　　要：详细分析了土壤背景偏振光谱二向反射分布函数与探测角及探测方位角之间关系.提出了基于最小二乘支持向量机回归的偏振光谱二向反射分布函数建模,将有限实验观测条件下测量得到的少数偏振光谱二向反射分布函数扩展到2π空间范围内任意入射及观测条件.通过模型结果和实验结果分析比较,表明该模型能很好地满足准确度要求.Relationship between detection geometry and spectropolarimetric bidirectional reflectance distribution function for soil background was analyzed in detail. The modeling of spectropolarimetric bidirectional reflectance distribution function was proposed based on LS-SVM, aiming to extend the value of measured spectropolarimetric bidirectional reflectance distribution function with limited geometry to whole hemi-sphere space. Results show that the predicted value and experimental value are close and meet the need of correction.

关 键 词：偏振光谱二向反射分布函数 探测几何条件 LS—SVM回归 

分 类 号：TN247[电子电信—物理电子学]