Abstract
This year’s plan is to build a database of tree species spectrum and stand structure characteristics.Eestablish a database of tree species spectral characteristics through hyperspectral images, extract hyperspectral information of different tree species in the ground sample area, extract spectral characteristic curves of each tree species, and establish tree species classification reference spectral characteristics, It also compares the applicable reference spectrum characteristics of the forest type and tree species, and uses the three-dimensional point cloud to produce forest structure indicators such as tree height, canopy coverage, and canopy penetration, and analyzes the structural differences of different tree species or stands.Develop an automated volume estimation model based on the AI algorithm, calculate the volume of the forest stand from the survey data of the ground sample area, and develop a multivariate estimation model with forest .Use the collected and built test forest area satellite telemetry image database, analyzing the difference in the accuracy of classification and interpretation of each sensor image application, and comparing it with the classification method of multi-time series and multi-scale fusion image information, and evaluating the integrated information of cross-scale satellite images Application potential of classification.