王成导师研究小组的论文在SCI期刊Remote Sensing Letters上发表

2020年9月24日，以云南师范大学云南省高校资源与环境遥感重点实验室2018级博士研究生施骏骋为第一作者、王成研究员为第二作者、习晓环副研究员为通讯作者的论文"Retrieving fPAR of Maize Canopy Using Artificial Neural Networks with Airborne LiDAR and Hyperspectral data"在SCI期刊Remote Sensing Letters在线发表（Remote Sensing Letters. 2020, Volume 11, Issue 11, doi: https://doi.org/10.1080/2150704X.2020.1807647）。该论文选择位于河北怀来县的中国科学院怀来遥感综合试验站玉米试验田为研究区，利用研究区的无人机激光雷达数据、无人机高光谱数据和地面实测数据，探索了LiDAR提取玉米冠层垂直结构信息、高光谱图像提取玉米光谱信息以及人工神经网络（ANNs）方法反演玉米光合有效辐射吸收率（fPAR）的潜力。首先，基于LiDAR点云数据提取了包括玉米高度和强度统计参数、冠层波动率、高度百分位数、累计高度百分位数、覆盖度等45个点云特征参数；然后，基于高光谱图像提取了NDVI、SRI、ARVI、RENDVI等13个特征参数；最后，使用ANNs和逐步多元线性回归（SMLR）方法构建了fPAR估算模型，并利用地面实测数据验证了模型的有效性。实验结果表明：结合两种数据构建的fPAR估算模型精度较单一数据有显著提升，且ANNs在fPAR反演中表现更好，可有效和可靠地估算玉米fPAR。LiDAR获取的冠层结构信息和高光谱数据提供的植被光谱特征可以实现互补，提高反演精度，这项研究为多源遥感技术精确估算玉米fPAR提供了参考，也进一步拓展了LiDAR在精准农业中的应用。

该论文得到了国家自然基金面上项目（41871264）和欧盟合作项目the Erasmus+ Capacity Building in Higher Education of the Education, Audiovisual and Culture Executive Agency (EACEA) for the “Innovation on Remote Sensing Education and Learning”(Grant no. 586037-EPP-1-2017-1-HU-EPPKA2-CBHE-JP)的联合资助。

论文相关信息

标题：Retrieving fPAR of Maize Canopy Using Artificial Neural Networks with Airborne LiDAR and Hyperspectral data

作者：Juncheng Shi ^{a, b, c}, Cheng Wang ^d, Xiaohuan Xi ^d*, Xuebo Yang ^d, Jinliang Wang ^{a, b, c}, Xue Ding ^{a, b, c}

通讯作者： Xiaohuan Xi, xixh@radi.ac.cn

作者单位：

a School of Tourism and Geographical Sciences, Yunnan Normal University, Kunming, Yunnan, CN;

b Key Laboratory of Resources and Environmental Remote Sensing for Universities in Yunnan, Kunming, Yunnan, CN;

c Center for Geospatial Information Engineering and Technology of Yunnan Province, Kunming, Yunnan, CN;

d Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, CN

出版物：Remote Sensing Letters

摘要：Accurate estimation of the fraction of absorbed photosynthetically active radiation (fPAR) is important for maize growth and yield estimations. Light detection and ranging (LiDAR)-derived canopy vertical structural and hyperspectral image-derived vegetation spectral information are complementary for vegetation fPAR estimation. This study explores the potential of artificial neural networks (ANNs) with two types of data to estimate maize fPAR. First, 45 metrics were derived from LiDAR data and 13 from a hyperspectral image. Then, the ANNs and stepwise multiple linear regression (SMLR) methods were used to estimate the fPAR. Finally, model validity was assessed using in-situ data. Results showed that the ANNs performed better in fPAR inversion (R² = 0.910, adj. R² = 0.921, RMSE = 0.046, RRMSE = 0.056, where R² is the coefficient of determination, adj. R² the adjusted coefficient of determination, RMSE the root mean squared error, and RRMSE the relative root mean squared error) than SMLR (R² = 0.638, adj. R² = 0.609, RMSE = 0.077, RRMSE = 0.092) and SMLR with the natural logarithm of data (R² = 0.855, adj. R² = 0.825, RMSE = 0.067, RRMSE = 0.081). This study is helpful for guiding the accurate estimation of maize fPAR using remote sensing techniques.

关键词：fPAR; airborne LiDAR; hyperspectral data; artificial neural networks; inversion