A Voxel-Based Fine-Scale 3D Landscape Pattern Analysis Using Point Clouds

The 3D landscape pattern metrics analysis method takes into account the structural information in the vertical direction of the landscape,analyzes the spatial heterogeneity of the landscape in 3D space,and solves the problem that the traditional 2D landscape pattern metrics analysis method cannot reflect the influence of the vertical structure of the landscape such as the topography and the height of landscape basic entity on landscape pattern characteristics.The rapid development of 3D space measurement technologies such as aerial oblique photogrammetry,Li DAR,and 3D laser scanning allows people to quickly obtain high-precision,high-resolution 3D landscape data,providing data support for the development of small-scale landscape 3D spatial pattern analysis methods.The voxel model describes the 3D structure of the object through 3D space pixels(voxels),which can finely describe the detailed features of the modeled object,and can build a high-resolution 3D landscape model combined with high-density point clouds,which has a wide application prospect in fine-scale 3D landscape pattern analysis.This paper constructed a high-resolution voxel model based on 3D laser scanner point clouds,constructed several voxel-based 3D landscape pattern metrics,and analyzed the 3D spatial heterogeneity of the landscape through the calculation of the3 D metrics and the comparison of the results of 2D and 3D metrics.The specific process is as follows:(1)Perform point cloud splicing,point cloud filtering,point cloud thinning,and point cloud classification processing on the original 3D point clouds collected by a 3D laser scanner to obtain classified point clouds.(2)Constructed high-resolution voxel model by voxelizing point clouds through the spatial grid division method,and evaluated the effect of the model construction.(3)Constructed composition and spatial configuration 3D landscape metrics based on voxel model to analyze the fine-scale 3D landscape spatial heterogeneity of the study area at class level and landscape level,and compared it with the real landscape pattern characteristics of the study area to verify the rationality of the 3D landscape metrics method based on voxels.(4)Compared the difference in the change trend of 2D and 3D calculation results of some landscape metrics,and analyzed the characteristics of the 3D spatial heterogeneity of the landscape.The results show that the voxel model constructed based on high-density point clouds has a high spatial resolution and a strong ability to describe landscape details,which can meet the requirements of fine-scale 3D spatial heterogeneity analysis of small-scale landscapes.The voxel model can reflect the phenomenon of vertical stratification of the landscape,and the description of the surface area,the relative position,and the pores of landscape basic physical are in good agreement with the real landscape.The voxel-based 3D landscape metrics method can comprehensively analyze the landscape composition and spatial configuration of the landscape in the 3D space from multiple aspects such as landscape composition,landscape fragmentation level,patch size and shape,patch compactness,and landscape vertical structure,and realize the correct description of the 3D spatial heterogeneity of the landscape.