Adaptive point-cloud filtering method for complex terrain structure

Abstract

The invention discloses an adaptive point-cloud filtering method for a complex terrain structure, and solves problems in the prior art that a filtering algorithm is sensitive to parameter change and misclassification is severe because of the fluctuant change of the terrain and the complex surface structure of an urban region. The method comprises the following steps: (1) building a point-cloud pyramid; (2) initializing a ground point G through employing the roughest hierarchical pyramid data, adding data of the next stage to a to-be-classified point U, and updating to-be-classified points; (3) carrying out the iteration of all hierarchies of the point-cloud pyramid, processing all the to-be-classified points, and storing a newly added ground point; (32) employing the method to obtain a filtering threshold value and carrying out filtering according to a DEM, bent energy and the scale information of the pyramid, and distinguishing the ground point and a non-ground point; (33) repeatedly carrying out steps (31)-(32) till there is no newly added ground point; (4) storing the ground point and the non-ground point after classification.

Description

technical field [0001] The invention belongs to the technical field of geospatial information systems, and mainly relates to a point cloud filtering method adaptive to complex terrain structures. Background technique [0002] With the rapid development of multi-platform lidar and multi-angle image dense matching technology, the high-resolution and high-precision terrain surface 3D reconstruction of a large number of quickly available point cloud data provides effective and reliable data support, point cloud data processing has become Hot issues in international research and industry. How to separate ground and non-ground points from point cloud data, that is, point cloud data filtering, is the first step in many subsequent applications, so it has always been a hot frontier issue. [0003] In actual production practice, due to the diversity and complexity of the surface structure, it is difficult for the same set of filtering parameters to adapt to the intricate terrain feat...

AI Technical Summary

Problems solved by technology

The main defects of this method include: on the one hand, terrain fluctuations and complex urban structures; on the other hand, the point cloud data generated by the dense matching of multi-platform lidar and multi-angle images has relatively poor point cloud density and data characteristics. However, the existing filtering algorithms usually use a set of parameters for the same area, which is difficult to adapt to the intricate ground-non-ground object structure, and is usually sensitive to parameter changes, resulting in serious misclassification problems

Although this method is highly efficient, it is difficult to deal with complex terrain point clouds because it is very sensitive to parameter selection, so it is not very practical.

Images