Research On Tree Parameter Extraction Of Masson Pine Forest Based On ALS And HLS Point Cloud Dat

Pinus massoniana is one of the main industrial wood species in our country and its economic value is high.It has become one of the main afforestation trees in southern China with its high comprehensive utilization and good timber characteristics.Laser radar has the characteristics of high resolution,good concealment and strong anti-interference ability.With the development of laser radar technology,the application of laser radar technology in the field of forestry is more and more extensive.Using Lidar to investigate forest resources,compared with traditional survey methods,this technology can obtain forest information quickly,economically,efficiently and in large area.When using airborne Lidar technology and handheld lidar to collect point cloud data,three-dimensional point cloud is obtained from different angles.Due to the complexity of forest structure,only airborne radar or handheld lidar technology is used to acquire forest information,which leads to the problem of insufficient acquisition of forest information.Therefore,in order to make up for the shortage of information acquisition by using a single lidar technology,this study took Masson Pine as the research object.Based on airborne lidar and handheld lidar,the Masson Pine forest point cloud data were obtained from the air and ground perspectives respectively.On the basis of preprocessing and fusion of ALS and HLS point cloud data obtained from the air and ground,By using the fused point cloud data,the segmentation algorithm of Masson pine forest and the extraction of tree parameters such as height,diameter at breast height and crown width of Masson pine forest were carried out,and the feasibility and effectiveness of the fusion of two kinds of data for the extraction of Masson pine tree parameters were discussed,in order to provide suitable monitoring methods and means for the rapid monitoring of Masson pine forest in the study area.The main research content and conclusions of this paper are summarized as follows:(1)This paper analyzes the current research status of point cloud single tree segmentation and parameter extraction at home and abroad;The principle of laser radar scanning technology is introduced.The composition and characteristics of lidar point cloud data are summarized.The theoretical knowledge of ALS and HLS point cloud data is summarized.This paper expounds the related technologies of multi-source data fusion,which lays a theoretical foundation for the follow-up research.(2)Aiming at the problem of incomplete information in point cloud data collected by single lidar technology,airborne lidar and handheld lidar were adopted in this paper to collect point cloud data of Masson Pine forest from the air and ground respectively.The registration strategy of "combining thickness and fine" was adopted to carry out the research on point cloud registration and fusion methods for the ALS and HLS point cloud data collected.In the registration fusion of ALS and HLS point cloud data,firstly,the Delaunay triangulation network was constructed to determine the corresponding points,and the rigid-body matrix transformation was used to complete the rough registration.Then,the iterative nearest point algorithm is used to complete the registration work.Finally,the fusion effect and accuracy were evaluated.Based on the measured data in the study area,the registration fusion experiment was carried out.The experimental results show that the registration fusion method can realize the high-precision fusion of point clouds with different data sources and different density levels.The mean square error of Pinus massoniana fusion was 0.058 m.(3)Based on the fused point cloud data,this paper carried out the single-wood segmentation algorithm suitable for Masson’s pine,and carried out the experimental comparative analysis and research of three point cloud segmentation algorithms based on PCS,CHM and layer stacking.Three point cloud segmentation algorithms were used to generate seed points,and then point cloud segmentation was carried out.Experimental results show that the relative error of seed points generated by PCS algorithm is 2.778% and the estimated accuracy is 97.222%.(4)At last,the paper carried out research on tree parameter extraction based on single tree segmentation,determined the process of tree parameter extraction based on single tree segmentation,extracted the height,DBH and crown width of single tree in Masson Pine forest,and analyzed the extraction accuracy.The results showed that the height of Masson pine could be effectively extracted by using the fusion point cloud data of ALS and HLS.There was a good linear relationship between the extracted tree parameter and the measured tree parameter.Compared with single point cloud data extraction tree height,fusion point cloud data extraction has obvious advantages.It can be seen that it is feasible and effective to extract tree parameters of Masson Pine forest based on ALS and HLS fusion point cloud data.