Research On Production Of Unmanned Aerial Vehicle(UAV) Image And Extraction Of Forest Information

Forest surveys is short for Forest resources surveys.The survey objects are the land,forests,plants and animals that live in the forest area,environmental factors.Forest surveys are surveys that aim at controlling the stock of forest resources and the growth of forests.In recent years,UAV have developed rapidly,sensors from RGB cameras,multi-spectral cameras,and laser radar have greatly promoted traditional forest survey methods.Not only reduces work intensity,but also improves work efficiency.This thesis mainly discusses the remote sensing image data based on UAV,uses the relevant UAV post-processing software to make geographic information products and restores the point cloud model of the research area;Research a method for multi-temporal images to realize high-precision automatic pricking control points based on existing DOM and DEM before AT process;Extract tree-measurement factors such as tree height,crown width,crown surface area and crown volume based on 4D products and 3D point cloud data,combine with ground survey data to verify the accuracy of forest information obtained by UAV.The main contents are as follows:(1)For the acquisition of multi-temporal forestland data,the basic geographical data updating,and high-resolution UAV images provide the basis for interpretation of remote sensing images.Using existing DOM and DEM,present a method without field work in laying control points,using IPS and auxiliary software to realize high-precision automatic pricking control points,which can be using for high-precision multi-scale matching.(2)Based on DOM,for large area of forest,when the display of individual trees on the images is similar,template matching methods are used for counting trees.For a regular pine forest,use the training sample to generate templates,and then use the template matching method to calculate the number of trees in the study area.(3)Based on image point cloud data,after denoising,normalization,and segmentation processing,extract single-tree information such as location,tree height,crown width,and crown volume.Then calculate the stand information like canopy density,average height,and density of the stands.Among them,the main innovations are:the use of photogrammetry software in the investigation of forest resources,the construction of simple point cloud data to extract forest stand information;Based on existing DOM,high resolution images and DEM,provide pricking control point method for different phases images before AT.