Three-dimensional Spatial Heterogeneity Analysis Of LiDAR Point Clouds For Crown Structures

With the progress of sciences and technology, laser products have become the "eyes" and "nervous". The laser radar remote sensing technology has an important role in the acquisition of the digital information. For ground objects, LiDAR point cloud data with3D coordinate information can represent the details of an object in space. Therefore it has been extensively applied in many fields.Crown structure is indispensable vital component of ecological process such as photosynthetic trees, researching its space structure has important significance. Due to the influence of external factors and of itself features, the spatial distribution of crown structure between different tree species or the same species exists some differences, this difference is the spatial heterogeneity. Anglicising spatial heterogeneity has many methods. Because of trees have fractal characteristics, fractal dimension is the original method. However, The ancestor of fractal theory, Mandelbrot found that spatial structure of objects with the same fractal dimension exists obvious difference. Using fractal dimension index to describe the differences in the spatial distribution isn’t enough, so other indicators (such as lacunarity index) to distinguish between different objects are required. Due to the importance of canopies, scholars have done a lot of research. Although there are many research results on the canopy, the related reports that using advantages of LiDAR point cloud data study crown structure of a tree doesn’t exist. Many studies are based on the optical image or the entire forest in the study area. Under the background of the above, this paper introduces the definition of lacunarity index and3D gliding-box algorithm. In the meantime, this research puts forward a lacunarity method based on three-dimensional convex hull and voxelization with the LiDAR point cloud data. This paper preliminary research lacunarity algorithm for a single crown and improves this method in forestry. The works carried out in this paper as follows:(1) We comprehensively overview state of the crown structure and its applications. Analyzed the main application field of the different data sources, methods and the main problems.(2) We describe LiDAR data and its advantages and disadvantages in detail.(3) The definition of lacunarity index and its original calculation method is described in detail and made some improvements on the basis of the original calculation method. At the same time, using the improved method to calculate the lacunarity index of tree crowns from point cloud data. Then, according to the experimental results, the effectiveness of this method is analysed.(4) Using three-dimensional shape signatures distinguished three-dimensional shapes of all tree crowns from LiDAR data and compared it with lacunarity index.(5) This paper has been viewed that lacunarity index can be a potential measure for classification of tree species using LiDAR data.In this paper, we measured the effectiveness of the improved algorithm of crown structures from LiDAR data. The experimental results show that LiDAR data are effective to study the distribution of tree crown in space. We provided data support for researching differences of the spatial distribution between tree crowns and ecological change process in forest management and developed a new method to distinguish the tree species.