确定SCS+C地形校正模型经验参数C的不同抽样方法的比较

Comparison of sampling approaches for determining empirical C-parameter of sun-canopy-sensor with C-correction topographic correction model

对于含有类似经验参数C的地形校正模型而言,确定经验参数的抽样方法是影响地形校正模型校正效果的一个关键因素。为此,比较分析了简单随机抽样、坡向分层抽样、坡度分层抽样和cos i分层抽样四种抽样方法分别计算的参数C的地形校正效果。基于cos i分层抽样和坡度分层抽样拟合的线性回归方程的相关系数R2比简单随机抽样和坡向分层抽样的高。实验表明,cos i分层抽样和坡度分层抽样确定的参数C的校正效果较好。

Sampling methods for deriving empirical parameters is an important factor influencing the performance of topographic correction models with the empirical parameters such as the parameter C.In order to examine the effectiveness of the C-parameters derived,this paper used a simple random sample,a sample stratified by aspect angles,a slope-stratified sample,and a sample stratified by cosines of solar incidence angles（i） to determine the parameters C of sun-canopy-sensor with C-correction（SCS＋C）.The R2 values of the linear correlation coefficients derived based on the cos i stratified and slope-stratified samples were higher than those derived based on the other two samples.The experimental results of this study show that the cos i stratified sample and the slope-stratified sample perform better than the other two sampling methods.