Study On Estimation Model And Quantitative Method Of Lvv Of Regional Vegetation Based On Fast 3D Reconstruction

As a new ecological quantitative index,living vegetation volume(LVV,also called three-dimensional green quantity)can reflect the change of greening effect with spatial distribution in the survey area and can be more directly related to other ecological indicators such as biomass.Thus,LVV is widely used in urban ecological environment assessment and urban planning.At the same time,three-dimensional(3D)point cloud data is widely used in the field of remote sensing and mapping,but it is less directly used in the estimation of LVV in urban areas.Among the sources of obtaining 3D point cloud data of urban vegetation,with the maturity of UAV remote sensing technology and 3D reconstruction technology,the real scene 3D reconstruction technology using UAV remote sensing images is evolving rapidly.This cross-field research provides a data acquisition means with faster response,lower cost,and richer ground object information for LVV estimation in urban areas.However,through literature review,we observe that the relevant research have the following problems:(1)The existing LVV quantitative model and methods are mainly based on the calculation of the growth parameters of a single tree species.Different groups of different model parameters are needed to deal with different types of tree species,so it is difficult to achieve the universality of the calculation model.(2)The existing LVV quantitative model or method is mainly based on single plant modeling,which cannot be applied to the situation of many tree species and complex forest structures in the region,so it is not suitable for large-scale application in urban areas.(3)The existing 3D reconstruction based on UAV remote sensing image takes too long and the efficiency is relatively low.In addition,due to the complex texture information of vegetation and other ground objects and the limited photographing angle of UAV remote sensing,the effect of vegetation reconstruction in the process of 3D reconstruction is poor and there is a lack of understory information,which makes it difficult to calculate the LVV of the urban area directly based on 3D point cloud data.Aiming at the above problems,this study investigates and designs the estimation model and quantitative method of LVV for regional vegetation based on the rapid 3D reconstruction of UAV remote sensing.The main research and original work include:(1)Using UAV tilt photogrammetry with a more comprehensive acquisition perspective,this study optimizes the algorithm process based on the principle of a 3D reconstruction algorithm.For the open-source 3D reconstruction algorithm,this study designs the optimization scheme of multi-core acceleration;for the commercial software of 3D reconstruction,this study presents the construction scheme of commercial software based on a parallel computing cluster.Thus,these two approaches jointly form the 3D point cloud rapid reconstruction technology based on UAV Remote Sensing to obtain the3 D point cloud in the study area and achieve the acceleration ratio of 6.88:1.(2)For the generated regional reconstruction point cloud,pre-processing is needed to extract the vegetation information of the study area before the calculation of LVV based on the 3D point cloud.Therefore,the vegetation point cloud extraction method based on deep learning and semi-automatic method are realized,and comparative experiments are carried out on a variety of point cloud data sets to determine the applicable scenes of different extraction methods in the accurate extraction of vegetation point cloud.(3)According to the characteristics of large differences in vegetation types and complex distribution in the scene,the point cloud modeling method based on the voxel model and the point cloud spatial organization and search method based on octree are proposed,and then the LVV modeling and calculation method of regional plants is designed through voxel space accumulation.On this basis,the problems existing in the calculation model are deeply analyzed.Combined with the inherent characteristics of the3 D reconstruction point cloud and the requirements of model universality,an improved LVV calculation method with scan line filling optimization is further proposed.In the experiment,the error rates of single plant and regional green quantity are 2.6% and 6.8%respectively,and the acceleration ratio of 200:1 is achieved compared with the traditional method.(4)Select a suitable site to carry out the experiment,take the traditional tree measurement factors such as crown height and crown diameter as the evaluation index,and use the traditional method to measure the accurate LVV to compare and verify the calculation results of the method proposed in this study.Taking the quadrat as the unit,calculate the local LVV value of each survey area in each test area,and generate the LVV spatial distribution thematic map of the survey area according to the LVV calculation results.The experimental results show that the LVV measurement method for urban area based on rapid 3D reconstruction of UAV remote sensing proposed in this study has obvious efficiency improvement and high measurement accuracy compared with the traditional method.Furthermore,this method can be applied to the scenario of single plant and multiple plants at the same time,which realizes the calculation of LVV of various types of vegetation in urban areas with high efficiency and high precision.