Grain yield prediction of rice using multi-temporal UAV-based RGB and multispectral images and model transfer – a case study of small farmlands in the South of China,Agricultural and Forest Meteorology

当前位置： X-MOL 学术 › Agric. For. Meteorol. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Grain yield prediction of rice using multi-temporal UAV-based RGB and multispectral images and model transfer – a case study of small farmlands in the South of China
Agricultural and Forest Meteorology ( IF 5.6 ) Pub Date : 2020-09-01 , DOI: 10.1016/j.agrformet.2020.108096
Liang Wan , Haiyan Cen , Jiangpeng Zhu , Jiafei Zhang , Yueming Zhu , Dawei Sun , Xiaoyue Du , Li Zhai , Haiyong Weng , Yijian Li , Xiaoran Li , Yidan Bao , Jianyao Shou , Yong He

Abstract Timely and accurate crop monitoring and yield forecasting before harvesting are valuable for precision management, policy and decision making, and marketing. The aim of this study is to explore the potential of fusing spectral and structural information extracted from the unmanned aerial vehicle (UAV)-based images in the whole growth period of rice to improve the grain yield prediction. A UAV platform carrying RGB and multispectral cameras was employed to collect high spatial resolution images of the rice crop under different nitrogen treatments over two years. The vegetation indices (VIs), canopy height and canopy coverage were extracted from UAV-based images, which were then used to develop random forest prediction models for grain yield. Among all of the investigated VIs, it was found that normalized difference yellowness index (NDYI) was the most useful index to monitor the changes in leaf chlorophyll content as well as the leaf greenness during the whole growth period. Meanwhile, the VIs provided a comparable prediction of grain yield to field-measured aboveground biomass and leaf chlorophyll content. Fusion of the multi-temporal normalized difference vegetation index (NDVI), NDYI, canopy height and canopy coverage achieved the best prediction of grain yield with a determination coefficient of 0.85 and 0.83, and relative root mean square error of 3.56% and 2.75% in 2017 and 2018, respectively, which outperformed the results in the reported studies. The initial heading stage was the optimal growth stage for the prediction of grain yield. Furthermore, the robustness of prediction model developed from the dataset in 2017 was validated by an external dataset from 2018 using model transfer. These findings demonstrate that the proposed approach can improve the prediction accuracy of grain yield as well as achieve an efficient monitoring of crop growth.

更新日期：2020-09-01

点击分享查看原文

点击收藏

阅读更多本刊最新论文本刊介绍/投稿指南11