王冬玲, 舒清态, 王强, 等. 基于偏最小二乘回归的高山松生物量遥感模型空间尺度效应分析[J]. 西南林业大学学报(自然科学), 2020, 40(4): 87–93 . DOI: 10.11929/j.swfu.201910021
引用本文: 王冬玲, 舒清态, 王强, 等. 基于偏最小二乘回归的高山松生物量遥感模型空间尺度效应分析[J]. 西南林业大学学报(自然科学), 2020, 40(4): 87–93 . DOI: 10.11929/j.swfu.201910021
Dongling Wang, Qingtai Shu, Qiang Wang, Hongbin Luo, Keren Wang. The Spatial Scale Effecton Analysis on Remote Sensing Estimation Model of Pinus densata Above-biomass Based on Partial Least Squares Regression[J]. Journal of Southwest Forestry University, 2020, 40(4): 87-93. DOI: 10.11929/j.swfu.201910021
Citation: Dongling Wang, Qingtai Shu, Qiang Wang, Hongbin Luo, Keren Wang. The Spatial Scale Effecton Analysis on Remote Sensing Estimation Model of Pinus densata Above-biomass Based on Partial Least Squares Regression[J]. Journal of Southwest Forestry University, 2020, 40(4): 87-93. DOI: 10.11929/j.swfu.201910021

基于偏最小二乘回归的高山松生物量遥感模型空间尺度效应分析

The Spatial Scale Effecton Analysis on Remote Sensing Estimation Model of Pinus densata Above-biomass Based on Partial Least Squares Regression

  • 摘要: 针对光学影像在森林生物量遥感估测中尺度不确定性问题,以香格里拉市高山松林为研究对象,基于SPOT-5影像,结合92块角规样地数据,在不同观测尺度下构建偏最小二乘回归模型,并对不同观测尺度下的遥感估测模型精度进行对比分析。结果表明:样地的合理观测尺度上限为70 m,生物量估测的最优空间尺度为50 m,此时PLS模型的RMSE最小为20.001 9 t/hm2R2和精度P最大,分别为0.505 8、66.20%。

     

    Abstract: Aiming at the uncertainty of scale of optical image in forest biomass remote sensing estimation. In this study, taking Pinus densata forest in Shangri-La as the research object, based on SPOT-5 image, combining with 92 sample plots of angle gauge measurement to build partial least squares regression models under different observation scales, and to analyzed the accuracy of models under different observation scales.The results show that the upper limit of the reasonable observation scale of the plot is 70 m, the optimal spatial scale for biomass estimation is 50 m, at this time, the RMSE of the PLS model is the smallest, which is 20.001 9 t/hm2, and R2, P reach the maximum, they are 0.505 8 and 66.20%, respectively.

     

/

返回文章
返回