Estimation Of Forest Parameter For Stand Volume In Coniferous Forest Based On Unmanned Aerial Vehicle Remote Sensing Images

Forest resource planning and design inventory is an important work of forestry activities.It is usually composed of different works in different industries.The aim is to understand the quantity,quality,distribution and change of forest resources in real time.With the development of computer technology and remote sensing technology,the intensity of field work in forest resources investigation is weakening,and the level of information and automation is improving.This has led to a significant improvement in the timeliness of forest resources monitoring by the forestry sector.As a remote sensing data collection platform,Unmanned Aerial Vehicle(UAV)has played a increasingly important role in forest resource survey.In southern subtropical forest area of China,due to the fragmentation,diversity and complex structure of forest,the technology of automatic extraction of forest information needs to be improved,and coniferous forest is the main forest ecosystem in subtropical area.Therefore,this paper select the coniferous forest plots in Jangle County in Fujian Province,which is located in the subtropical region as the study area,The demand for estimation of forest storage volume is based on the data of UAV remote sensing as the main data source,supplemented by airborne LiDAR and other data.According to the different types of forestry plots,extracting the information of the number of trees,tree canopy and tree height,the volume of forest sub-compartment was estimated.The results of the study are compared with the forest sub-compartment and field data.The main research contents and conclusions are as follows:(1)Analysis of the preprocessing of UAV:The images of UAV in forest area are uniform exposure and high overlap.With the help of Pix4D processing platform,the high accuracy and high resolution digital surface model(DSM)and digital orthophoto map(DOM)were generated by UAV images.The true color Orthophoto image was natural and no obvious stitching line.(2)Extraction and evaluation of Coniferous forest stand parameters based on single period UAV images.The ability to extract stand parameters form single period UAV images was analyzed.The results showed that the single period UAV images can effectively extract the canopy,canopy density and tree number,the accuracy of the estimation is 90%,but the ability of tree height estimation is limited.The single period UAV images can only extract the tree high in the low canopy density area,and the terrain information can not be obtained in the middle and high canopy density area.(3)Estimation of forest stand volume on forest felling area based on two periods UAV images.According to the demand of forest stand volume estimation,this paper selected pre-harvest and post-harvest UAV images to carry out the estimation of forest cutting volume.The methods of the two periods point cloud matching and point cloud classification by cloth simulation filtering algorithm were analyzed.The effect of extracting canopy vertices from canopy height mode using adaptive local maximum algorithm was evaluated.A random forest model was established by the measured data of tree height and single wood volume which is collected in the field,the forest stand volume on forest felling area was effectively estimated,the accuracy of the estimation is 97%.(4)Estimation of coniferous sub-compartment stem volume based on airborne LiDAR and UAV images.In order to estimate coniferous sub-compartment stem volume in the 2 square kilometers range,the data sources used UAV data and airborne LiDAR data.The results showed that the progressive trigonometric filtering algorithm can effectively extract the topographic points in the LiDAR point cloud and acquire high precision DEM data.The UAV remote sensing data is used to produce canopy height model with the help of DEM data.Based on this,the average tree height of small classes in the study area was extracted.At the same time,with the help of CHM thematic feature to extract canopy and tree number information,the volume of large area coniferous forest in forest sub-compartment was quickly extracted.