Leaf Nitrogen Concentration Retrieval Based on Polarization Reflectance Model and Random Forest Regression
ZHANG Zi-han1, YAN Lei1,2, LIU Si-yuan1, FU Yu1, JIANG Kai-wen1, YANG Bin3, LIU Sui-hua4, ZHANG Fei-zhou1*
1. Beijing Key Lab of Spatial Information Integration and 3S Application, Institute of Remote Sensing and Geographic Information System, School of Earth and Space Science, Peking University, Beijing 100871, China
2. Guangxi Key Lab of UAV Remote Sensing, Guilin University of Aerospace Technology, Guilin 541004, China
3. College of Electrical and Information Engineering, Hunan University, Changsha 410082, China
4. School of Geography and Environmental Science, Guizhou Normal University, Guiyang 550001, China
Abstract:Leaf nitrogen concentration is of great significance in the vegetation biochemistry process. Airborne hyperspectral data is widely utilized to retrieve leaf nitrogen concentration. Since the current algorithms cannot completely fulfill the accuracy requirement of precision agriculture, it is urgent to improve the retrieval accuracy of leaf nitrogen concentration. The accuracy of leaf nitrogen concentration retrieval is restricted by principle error and algorithm error. The principle error is generated in the process of specular reflection at the leaf surface. The radiant energy detected by sensors consists of a specular components and multiple scattering components. Solely the multiple scattering component carries vegetation biochemistry information (leaf nitrogen concentration, for instance). The specular component represents the energy reflected directly at the foliar wax layer, thus carries no inner information of the leaf. Based on the Fresnel formula, the specular component is partially polarized, and the multiple scattering component is unpolarized. Therefore, the principle error can be eliminated by the specular reflectance estimate, particularly with the aid of polarization reflectance modelling. The algorithm error is derived from the difference of airborne hyperspectral data mining capability between different algorithms. The performance of Partial Least Squares Regression, Principal Component Regression, Support Vector Regression, K-Nearest Neighbor Regression and Random Forest Regression are systematically compared in this research, and ultimately Random Forest Regression is chosen to reduce the algorithm error. In this research, in order to estimate the polarization reflectance of broadleaf and needle vegetation, multispectral data gained by POLDER/PARASOL satellite (equipped with multi-angle polarization sensors) are used to establish Bidirectional Polarization Distribution Function model. Hyperspectral data gained by the HySpex sensor system is used to conduct high-precision retrieval of leaf nitrogen concentration. Root Mean Square Error is taken as a major evaluation index. The conclusion is: After eliminating polarization reflectance in hyperspectral data, an average accuracy improvement of 4.244% is achieved among the above algorithms. Random Forest Regression is rather competitive by reaching 13.103% improvement in accuracy (RSQ 0.803, RMSE 0.252), which indicates that Random Forest is sensitive to polarization information. Compared to the basic method (Partial Least Squares Regression), the accuracy is improved by 32.440% after eliminating principle error and reducing algorithm error. In our research, the high-accuracy retrieval of leaf nitrogen concentration is realized, proving the significance of eliminating polarization reflectance and indicates the potential of random forest regression in hyperspectral remote sensing retrieval.
Key words:Remote sensing retrieval; Polarization remote sensing; Leaf nitrogen concentration; Hyperspectral data; Random forest regression; Bidirectional polarization distribution function
张子晗,晏 磊,刘思远,付 瑜,姜凯文,杨 彬,刘绥华,张飞舟. 基于偏振反射模型和随机森林回归的叶片氮含量反演[J]. 光谱学与光谱分析, 2021, 41(09): 2911-2917.
ZHANG Zi-han, YAN Lei, LIU Si-yuan, FU Yu, JIANG Kai-wen, YANG Bin, LIU Sui-hua, ZHANG Fei-zhou. Leaf Nitrogen Concentration Retrieval Based on Polarization Reflectance Model and Random Forest Regression. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2021, 41(09): 2911-2917.
[1] Kokaly R F, Asner G P, Ollinger S V, et al. Remote Sensing of Environment, 2009, 113: S78.
[2] Li F, Mistele B, Hu Y, et al. European Journal of Agronomy, 2014, 52: 198.
[3] Yao X, Zhu Y, Tian Y, et al. International Journal of Applied Earth Observation and Geoinformation, 2010, 12(2): 89.
[4] Chlingaryan A, Sukkarieh S, Whelan B. Computers and Electronics in Agriculture, 2018, 151: 61.
[5] Yao X, Huang Y, Shang G, et al. Remote Sensing, 2015, 7(11): 14939.
[6] Yang B, Zhao H, Chen W. Journal of Quantitative Spectroscopy and Radiative Transfer, 2017, 202: 13.
[7] Wang Z, Skidmore A K, Wang T, et al. International Journal of Applied Earth Observation and Geoinformation, 2017, 54: 84.
[8] Bréon F M, Maignan F. Earth System Science Data, 2017, 9(1): 31.
[9] Knyazikhin Y, Martonchik J V, Myneni R B, et al. Journal of Geophysical Research: Atmospheres, 1998, 103(D24): 32257.
[10] Nadal F, Bréon F M. IEEE Transactions on Geoscience and Remote Sensing, 1999, 37(3): 1709.
[11] Waquet F, Leon J F, Cairns B, et al. Applied Optics, 2009, 48(6): 1228.
[12] Maignan F, Bréon F M, Fédèle E, et al. Remote Sensing of Environment, 2009, 113(12): 2642.
[13] Litvinov P, Hasekamp O, Cairns B. Remote Sensing of Environment, 2011, 115(2): 781.
[14] Diner D, Xu F, Martonchik J, et al. Atmosphere, 2012, 3(4): 591.