Research on Classification of Pest and Disease Tree Samples Based on Hyperspectral Lidar

In this study, a set of tunable hyperspectral lidar system with 91 channels, spectral resolution of 5 nm, wavelength range of 650-1100 nm, and high biological safety is built, and the detection experiments of forest tree samples such as Ailanthus altissima, Pinus yunnanensis, and Koelreuteria paniculata are completed. The target echo intensity is detected through experiments, and the target spectral reflectance is obtained. Finally, the support vector machine classifier is used to classify and identify different types of healthy and diseased samples. The classification accuracy of Ailanthus altissima samples can reach 96. 98%. The classification accuracy of Pinus yunnanensis samples can reach 91.21%, and the classification accuracy of Koelreuteria paniculata samples can reach 66. 21%. The experimental results have research significance and reference value, and provide a new development direction for forestry pest monitoring.