外文摘要：Timely acquisition of soil moisture content (SMC) in the root zone of field crops is crucial for achieving precise irrigation. In this study, high-spectral technology was employed to collect both SMC data at different soil depths during the bolting stage of winter oilseed rape (Brassica napus L.) and corresponding hyperspectral data through two consecutive years of field experiments (2021-2022). Two categories of spectral parameters were established, including "three-edge" spectral parameters such as blue, yellow, and red edge areas, as well as empirical vegetation indices that showed strong correlations with previous research and crop parameters. The spectral parameters with the highest correlation coefficients with SMC at various soil depths were selected. Subsequently, the selected vegetation indices were utilized as inputs for modeling. Support vector machine (SVM), random forest (RF), back propagation neural network (BPNN), and extreme learning machine (ELM) were employed to construct SMC estimation models for different soil depths during the jointing stage of winter oilseed rape. The results indicated that the majority of the three-edge parameters and empirical vegetation indices displayed higher correlation coefficients with SMC in the 0-20 cm soil depth compared to the 20-40 cm and 40-60 cm depths. The RF model proved to be the optimal modeling method for SMC estimation, particularly for the 0-20 cm soil depth. The validation set of the estimation model for this depth achieved a high coefficient of determination (R-2) of 0.944, a root-mean-square error (RMSE) of 0.005, and a mean relative error (MRE) of 3.074%. The findings of this study provide a basis for high-spectral monitoring of SMC in the root zone of winter oilseed rape, offering valuable insights for the rapid assessment of crop growth under water stress conditions.
外文关键词：hyperspectral;machine learning;soil moisture content;Winter oilseed rape
作者：Zhang, Wei;Zhang, Fucang;Wang, Xin;Tang, Zijun;Xiang, Youzhen;Liu, Xiaochi;Shi, Hongzhao;Li, Zhijun
作者单位：Northwest A&F Univ
期刊名称：JOURNAL OF SOIL SCIENCE AND PLANT NUTRITION
期刊影响因子：0.0
出版年份：2024
出版刊次：24（1）
原文传递申请：江苏省科技资源（工程技术文献）统筹服务平台