remote sensing Article Retrieving Soybean Leaf Area Index from Unmanned Aerial Vehicle Hyperspectral Remote Sensing: Analysis of RF, ANN, and SVM Regression Models Huanhuan Yuan 1,2,3 , Guijun Yang 1,3,4, *, Changchun Li 2 , Yanjie Wang 1,2 , Jiangang Liu 1,3 , Haiyang Yu 1,4 , Haikuan Feng 1,3 , Bo Xu 1,4 , Xiaoqing Zhao 1,4 and Xiaodong Yang 1,3,4 1 Beijing Research Center for Information Technology in Agriculture, Key Laboratory of Quantitative Remote Sensing in Agriculture, Ministry of Agriculture, Beijing 100097, China; yuanhuanhuan199@163.com (H.Y.); wangyj.gmai@gmail.com (Y.W.); ljgwr0619@sina.com (J.L.); yuhy@nercita.org.cn (H.Y.); fenghk@nercita.org.cn (H.F.); xub@nercita.org.cn (B.X.); zhaoxq@nercita.org.cn (X.Z.); yangxd@nercita.org.cn (X.Y.) 2 School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, China; lichangchun610@126.com 3 National Engineering Research Center for Information Technology in Agriculture, Beijing 100097, China 4 Beijing Engineering Research Center for Agriculture Internet of Things, Beijing 100097, China * Correspondence: guijun.yang@163.com; Tel.: +86-10-51-503-647; Fax: +86-10-51-503-750 Academic Editors: Zhenhong Li, Clement Atzberger and Prasad S. Thenkabail Received: 28 December 2016; Accepted: 21 March 2017; Published: 25 March 2017 Abstract: Leaf area index (LAI) is an important indicator of plant growth and yield that can be monitored by remote sensing. Several models were constructed using datasets derived from SRS and STR sampling methods to determine the optimal model for soybean (multiple strains) LAI inversion for the whole crop growth period and a single growth period. Random forest (RF), artificial neural network (ANN), and support vector machine (SVM) regression models were compared with a partial least-squares regression (PLS) model. The RF model yielded the highest precision, accuracy, and stability with V-R 2 , SD R 2 , V-RMSE, and SD RMSE values of 0.741, 0.031, 0.106, and 0.005, respectively, over the whole growth period based on STR sampling. The ANN model had the highest precision, accuracy, and stability (0.452, 0.132, 0.086, and 0.009, respectively) over a single growth phase based on STR sampling. The precision, accuracy, and stability of the RF, ANN, and SVM models were improved by inclusion of STR sampling. The RF model is suitable for estimating LAI when sample plots and variation are relatively large (i.e., the whole growth period or more than one growth period). The ANN model is more appropriate for estimating LAI when sample plots and variation are relatively low (i.e., a single growth period). Keywords: LAI retrieval; hyperspectral remote sensing; sampling method; random forests; artificial neural networks; support vector machine 1. Introduction Soybeans are the most widely grown oil crops in the world. Soybean leaf area index (LAI) reflects photosynthetic rate [1,2] and crop yield [3]. Therefore, LAI is an essential parameter for breeding high-yield soybean plants [4,5]. Methods for estimating LAI can be categorized as direct methods and indirect methods [6,7]. Indirect methods adopt devices such as plant canopy analyzers (e.g., LAI-2000, LI-COR, Inc., Lincoln, NE, USA), digital hemispherical photography (DHP), and remote sensing [8]. Remote sensing technology is cost-effective and non-destructive and, thus, a prevalent technology for estimating LAI [9]. Remote Sens. 2017, 9, 309; doi:10.3390/rs9040309 www.mdpi.com/journal/remotesensing