Abstract:

Unmanned aerial vehicle(UAV) has been widely used in forest resource investigation because of its low cost and high precision.The study on estimation and evaluation methods of main forest parameters using UAV high-resolution image and point cloud data can provide scientific reference for the spread and application of UAV technology in plantation forest survey.In this paper,experimental plantation forest of Taxodium mucronatum in the arboretum of Nanjing Forestry University was selected as the research object.The data of high-resolution image of UAV,point cloud and ground observation in 2018 were used as the main information sources.The number of trees,tree height,crown width and canopy closure were calculated by local maximum and seed point segmentation.The parameters are extracted and the accuracy is evaluated.The results show that:(1) the detection rate of the extracted tree is 0.92,the accuracy of the number of trees is 0.97,and the value of F parameter is 0.95;(2) the determination coefficient R² of single tree height is 0.7957,the root mean square error RMSE of height is 0.5940,and the determination coefficient R² of crown diameter is 0.8008,and the root mean square error RMSE of diameter is 0.8978.(3) The extraction rate of total crown area was 0.95,the accuracy rate was 0.93,and the F parameter was 0.94;(4) The relative error of canopy density precision was only 0.32 percent.Estimation of main parameters of plantation using UAV high spatial resolution data and a small amount of field survey data can replace the actual field survey of the whole forest,and has great application value in plantation forest survey.

Key words: plantation forest, forest parameters, UAV(Unmanned aerial vehicle), taxodium mucronatum, local maximum method