Litchi disease detection method and system based on hyperspectrum of unmanned aerial vehicle

The invention discloses a litchi disease detection method and system based on unmanned aerial vehicle hyperspectrum. Local hyperspectral data and visible light data are obtained and processed through an airborne hyperspectral camera carried by an unmanned aerial vehicle; processing and splicing the local hyperspectral data into a hyperspectral stereoscopic panorama; and selecting an area with a specific size, inputting the spectral data in the area into a pre-constructed disease data model, and judging the plant health condition and the disease type in the area. According to the method, a depth image splicing model and a disease data model are constructed by using 2D and 3D convolutional neural networks, the design of a residual structure enables network operation to be simpler, the required calculation amount is reduced, and data support is provided for large-scale orchard disease divergence trend, disease monitoring and early warning by using a hyperspectral remote sensing technology. 本发明公开了一种基于无人机高光谱的荔枝病害检