3D Reconstruction Of Fruit Trees In Orchards Based On Laser Point Cloud Data

For the analysis and management of fruit tree growth,the collection of data about fruit trees in plantations is crucial.With the development of three-dimensional ground-based laser scanning technology equipment,three-dimensional laser point cloud to accomplish data acquisition and visualization model research of fruit tree commodities is becoming an increasingly popular area of forestry informatization.As a result,technological advancement has enhanced the informatization and intelligence of forestry.Therefore,in this paper,the data of apple trees in orchards collected by multi-station ground-based LIDAR is used as the research object to extract the data information of fruits and establish a 3D model,to separate the branches and leaves of fruit trees,to establish the branch model of fruit trees,and to visualize the fruit tree data information in 3D.The following sections comprise the specific content:(1)Program design and data acquisition.It consists primarily of the following steps:scanning site design,data collection,preprocessing,etc.Initially,the scanning site design of multi-site scanning was planned,and the point cloud data of fruit trees in orchards were scanned at different times using a Faro Fcous scanner;the point cloud alignment and data format transformation were performed using Scene,the scanner’s supporting processing software;on this basis,statistical filtering was used to perform noise reduction processing on the point cloud data;the separation processing of the ground surface from the point cloud data was also performed;and finally,the separation processing of the ground surface from the On this basis,statistical filtering is employed to reduce the noise in the point cloud data;CSF filtering is employed to separate the ground point cloud in order to extract the fruit point cloud data.(2)Extraction and reconstruction of apple fruit restoration point cloud.In this paper,we use the RANSAC algorithm to extract the spherical point cloud of apple fruit,combine the biological morphological size of the fruit and set the threshold classification method to filter out the real apple fruit point cloud,and then fit it to obtain the size data and spatial location information.This paper proposes a spatial coordinate system applicable to a single fruit tree as the object,and the spatial distribution of fruits is more clearly visible;for the 3D model of fruits,it adopts the apple approximation model,and achieves the restoration of spherical fruits by proposing the restoration parameters based on the apple fruit shape index,and the restored apple frui The produce volume erroneous value is below 15 cm~3.The experimentally obtained data on apple fruit volume and spatial distribution can provide more precise data reference information for fruit information management and yield estimation and prediction.(3)Branch and foliage separation of point clouds for fruit trees.In recent years,the LeWoS model is a novel model for branch and leaf separation of solitary trees.Additionally,we restore and reconstruct the point cloud in 3D to obtain a 3D volume model of the fruit tree branches.