Airborne Hyperspectral Imaging in Early Monitoring of Pine Wood Nematode
OPTOSKYBLOGBlogsHyperspec Blog
Airborne Hyperspectral Imaging in Early Monitoring of Pine Wood Nematode
Airborne Hyperspectral Imaging in Early Monitoring of Pine Wood Nematode
author: Gavin
2022-01-14

Airborne Hyperspectral Imaging in Early Monitoring of Pine Wood Nematode
Pine wood nematode (PWD, the "cancer" of pine trees) is one of the devastating international forest diseases caused by the pine wood nematode (PWN; Bursaphelenchus xylophilus), which has caused huge ecological and economic losses to our country. An effective way to prevent large outbreaks of PWD is to detect and remove damaged pine trees in the early stages of PWD infection. However, in the visible wavelength range, early-infected pines fail to show obvious morphological or color changes, making early detection of PWD difficult.
Principle:
UAV-based hyperspectral imaging has great potential for early detection of PWD. Commonly used methods such as three-dimensional convolutional neural networks (3D-CNN) are able to simultaneously extract and fully exploit spatial and spectral information from raw spectral data.
Solution:
ATHL9010 series is the 3rd generation hyperspectral imager integrated LiDAR system.Use ATHL9010 to conduct field experiments to obtain hyperspectral data and radar data.

Reflectance curves for hardwood, early infected pine, and late infected pine.

The recognition and classification results of PWD by different models.
Table 4 Classification accuracy of 3 tree categories by different methods.


The effect of different training sample sizes on the accuracy of 3D-Res CNN models.
Pine wood nematode (PWD, the "cancer" of pine trees) is one of the devastating international forest diseases caused by the pine wood nematode (PWN; Bursaphelenchus xylophilus), which has caused huge ecological and economic losses to our country. An effective way to prevent large outbreaks of PWD is to detect and remove damaged pine trees in the early stages of PWD infection. However, in the visible wavelength range, early-infected pines fail to show obvious morphological or color changes, making early detection of PWD difficult.
Principle:
UAV-based hyperspectral imaging has great potential for early detection of PWD. Commonly used methods such as three-dimensional convolutional neural networks (3D-CNN) are able to simultaneously extract and fully exploit spatial and spectral information from raw spectral data.
Solution:
ATHL9010 series is the 3rd generation hyperspectral imager integrated LiDAR system.Use ATHL9010 to conduct field experiments to obtain hyperspectral data and radar data.

Result:

Reflectance curves for hardwood, early infected pine, and late infected pine.

The recognition and classification results of PWD by different models.
Table 4 Classification accuracy of 3 tree categories by different methods.


The effect of different training sample sizes on the accuracy of 3D-Res CNN models.
Conclusion:
In this study, the authors applied residual blocks to 3D CNN based on hyperspectral images and constructed a 3D-Res CNN model for early detection of PWD. Furthermore, the classification accuracy of 2D-CNN, 3D-CNN, 2D-Res CNN, and 3D-Res CNN models for identifying PWD-infected pine trees was compared. The results show that the 3D-Res CNN model is the most effective. It can extract spectral information and spatial information from hyperspectral images at the same time. Although the training time is long, it can still meet the needs of large-scale practical forestry applications.
Related Products:
Drone Hyperspectral Imaging System:ATHL9010
Airborne Hyperspectral Remote Sensing System:ATH9010
In this study, the authors applied residual blocks to 3D CNN based on hyperspectral images and constructed a 3D-Res CNN model for early detection of PWD. Furthermore, the classification accuracy of 2D-CNN, 3D-CNN, 2D-Res CNN, and 3D-Res CNN models for identifying PWD-infected pine trees was compared. The results show that the 3D-Res CNN model is the most effective. It can extract spectral information and spatial information from hyperspectral images at the same time. Although the training time is long, it can still meet the needs of large-scale practical forestry applications.
Related Products:
Drone Hyperspectral Imaging System:ATHL9010
Airborne Hyperspectral Remote Sensing System:ATH9010
Related News
Precision fertilization by UAV for rice at tillering stage in cold region based on hyperspectral remote sensing prescription map
2023-02-01 117Application of UAV-hyperspectral imaging for rice growth monitoring
2023-01-18 161Application of hyperspectral camera in industrial inspection
2023-01-17 109Estimation Scheme of Rice Yield Based on UAV Hyperspectral Images
2023-01-17 80Surveillance of pine wood nematode disease based on satellite remote sensing images
2023-01-16 78UAV forest fire patrol protection plan
2023-01-16 124Application of Hyperspectral Imager in Disguised Target Recognition(part 2)
2023-01-13 104Application of hyperspectral imager in detection of exogenous pests in jujube fruit
2023-01-11 126