文献类型： 外文期刊

作者： Zhao, Dan ¹ ; Yang, Hao ² ; Yang, Guijun ² ; Yu, Fenghua ¹ ; Zhang, Chengjian ³ ; Chen, Riqiang ³ ; Tang, Aohua ⁴ ; Zhang, Wenjie ³ ; Yang, Chen ⁵ ; Xu, Tongyu ¹ ;

作者机构： 1.Shenyang Agr Univ, Sch Informat & Elect Engn, Shenyang 110866, Peoples R China

2.Beijing Acad Agr & Forestry Sci, Informat Technol Res Ctr, Key Lab Quantitat Remote Sensing Agr Minist Agr &, Minist Agr & Rural Affairs, Beijing 100097, Peoples R China

3.Beijing Forestry Univ, Sch Informat Sci & Technol, Beijing 100083, Peoples R China

4.Changan Univ, Coll Geol Engn & Geomat, Xian 710064, Peoples R China

5.Henan Polytech Univ, Sch Surveying Mapping & Land Informat Engn, Jiaozuo 454000, Peoples R China

关键词： maize above-ground biomass; convolutional neural network; transfer learning; allometric growth model; UAV multispectral

期刊名称：REMOTE SENSING （ 影响因子：4.1； 五年影响因子：4.8 ）

ISSN：

年卷期： 2024 年 16 卷 16 期

页码：

收录情况： SCI

摘要： The precise estimation of above-ground biomass (AGB) is imperative for the advancement of breeding programs. Optical variables, such as vegetation indices (VI), have been extensively employed in monitoring AGB. However, the limited robustness of inversion models remains a significant impediment to the widespread application of UAV-based multispectral remote sensing in AGB inversion. In this study, a novel stem-leaf separation strategy for AGB estimation is delineated. Convolutional neural network (CNN) and transfer learning (TL) methodologies are integrated to estimate leaf biomass (LGB) across multiple growth stages, followed by the development of an allometric growth model for estimating stem biomass (SGB). To enhance the precision of LGB inversion, the large-scale remote sensing data and image simulation framework over heterogeneous scenes (LESS) model, which is a three-dimensional (3D) radiative transfer model (RTM), was utilized to simulate a more extensive canopy spectral dataset, characterized by a broad distribution of canopy spectra. The CNN model was pre-trained in order to gain prior knowledge, and this knowledge was transferred to a re-trained model with a subset of field-observed samples. Finally, the allometric growth model was utilized to estimate SGB across various growth stages. To further validate the generalizability, transferability, and predictive capability of the proposed method, field samples from 2022 and 2023 were employed as target tasks. The results demonstrated that the 3D RTM + CNN + TL method outperformed best in LGB estimation, achieving an R-2 of 0.73 and an RMSE of 72.5 g/m(2) for the 2022 dataset, and an R-2 of 0.84 and an RMSE of 56.4 g/m(2) for the 2023 dataset. In contrast, the PROSAIL method yielded an R-2 of 0.45 and an RMSE of 134.55 g/m(2) for the 2022 dataset, and an R-2 of 0.74 and an RMSE of 61.84 g/m(2) for the 2023 dataset. The accuracy of LGB inversion was poor when using only field-measured samples to train a CNN model without simulated data, with R-2 values of 0.30 and 0.74. Overall, learning prior knowledge from the simulated dataset and transferring it to a new model significantly enhanced LGB estimation accuracy and model generalization. Additionally, the allometric growth model's estimation of SGB resulted in an accuracy of 0.87 and 120.87 g/m(2) for the 2022 dataset, and 0.74 and 86.87 g/m(2) for the 2023 dataset, exhibiting satisfactory results. Separate estimation of both LGB and SGB based on stem and leaf separation strategies yielded promising results. This method can be extended to the monitor and inversion of other critical variables.