PDF
Full
Outline
Rubber is a globally significant strategic raw material, playing an irreplaceable role in national defense, transportation, aerospace and medical equipment industries. However, pests such as the rubber tree leaf mite frequently occur in tropical rubber plantations, especially under conditions of high temperature and drought, where they are prone to large-scale outbreaks. Infestations result in symptoms such as leaf chlorosis, yellowing and premature defoliation, which severely compromise tree vitality and substantially reduce latex yield. Traditional pest and disease monitoring relies primarily on manual field surveys, which are time-consuming, labor-intensive, and limited in spatial scope. The constraints make it increasingly difficult to meet the needs of modern rubber plantations for rapid early warning and precise pest management. Remote sensing technology has demonstrated significant advantages in the monitoring of agricultural pests and diseases. It enables large-scale, rapid, non-contact and dynamic monitoring across multiple temporal scales. To address the limitations of conventional monitoring methods, such as low efficiency and inadequate accuracy, this study proposes a multi-source remote sensing approach that integrates ground-based ASD (Analytical Spectral Devices) hyperspectral data, proximal multispectral imagery, and field survey data to achieve rapid and accurate detection of rubber tree mite infestations. The study was conducted based on spectral reflectance data of rubber leaves collected in 2024 from the experimental station of Rubber Research Institute, Chinese Academy of Tropical Agricultural Sciences in Danzhou, Hainan province, along with UAV-based multispectral imagery. The raw spectral data were preprocessed using Multiplicative Scatter Correction (MSC), Savitzky-Golay (SG) smoothing and first-order derivative transformation. Four types of spectral feature variables were derived, including vegetation indices (VIs), hyperspectral features, continuum removal (CR) parameters and wavelet coefficients. Sensitive features were selected through a combination of the Least Absolute Shrinkage and Selection Operator (LASSO) and Pearson correlation coefficient analysis. Based on the selected features, three regression models, Multiple Linear Regression (MLR), Random Forest (RF) and Back Propagation Neural Network (BPNN), were developed to model rubber tree mite infestations levels. Results indicated that the ASD hyperspectral data yielded the best performance across all models, with the RF and BPNN models performing particularly well. Among them, the RF model achieved the highest coefficient of determination (R²) of 0.8361 and the lowest root mean square error (RMSE) of 0.6353 on the validation dataset, demonstrating a strong capacity for accurate infestation severity prediction. In conclusion, the integration of ASD hyperspectral data with the RF model enables the construction of a high-accuracy, strongly interpretable monitoring model for rubber tree mite infestations. This approach effectively supports early detection and quantitative classification of infestation levels in rubber tree mite infestations, and would provide essential technical guidance and theoretical support for precision plantation management and integrated pest control.
PDF
43
13
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62199257
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
