Unmanned aerial vehicle hyperspectral image pine wood nematode disease identification method based on deep learning

Abstract

The invention provides an unmanned aerial vehicle hyperspectral image pine wood nematode disease identification method based on deep learning. The method comprises the following steps: S1, extracting a suspected diseased wood area from a hyperspectral image obtained by an unmanned aerial vehicle through an image segmentation network; S2, extracting a spectral feature vector and a spatial feature vector of the suspected diseased wood area through a deep learning network; S3, identifying the category of the suspected diseased wood area according to the spectral feature vector and the spatial feature vector. According to the method, autumn color-changing larch trees, bare forest lands and pine wood nematode disease wood can be effectively distinguished, the disease wood can be recognized with high accuracy even for low-canopy-density forest areas and in the middle and late periods of autumn, and the timely and accurate disease wood monitoring effect can be achieved.

Description

technical field [0001] The invention belongs to the technical field of monitoring forestry diseases and insect pests, and in particular relates to a method for identifying pine wood nematode disease based on hyperspectral images of unmanned aerial vehicles based on deep learning. Background technique [0002] Pine wood nematode is a devastating pine tree disease caused by pine wood nematode, and pine wood nematode is currently the most serious disease and insect pest that causes the loss of forest resources in my country. UAV remote sensing can quickly collect high-resolution images of large-scale key forest areas, and timely obtain information on the location of individual trees affected by pine wood nematode disease. It has good monitoring effects and great potential. The outbreak of pine wood nematode disease occurs in autumn every year, but the current monitoring of pine wood nematode disease with UAV visible light images or multi-spectral remote sensing images cannot ef...

AI Technical Summary

Problems solved by technology

The outbreak of pine wood nematode disease occurs in autumn every year, but the current monitoring of pine wood nematode disease with UAV visible light images or multi-spectral remote sensing images cannot effectively distinguish the color-changing deciduous trees in autumn, forest bare land, and pine wood nematode-affected trees. Monitoring in forest areas with high canopy density in early autumn, while the monitoring effect in forest areas with low canopy density and in middle and late autumn is poor

Images