When capturing an object using laser scanning, multiple viewpoints are utilised due to the recording perspective and required overlap for registration. During various R&D projects, methods were developed for the interactive and automated evaluation of merged point clouds in metigo® software. Pre-import of registered viewpoints allows loading all data in reduced quality to define a relevant object area, leading to significant initial data reduction. Individual viewpoints are merged into an overall point cloud with specified point spacing. A voxel grid, consisting of cube-shaped cells, is created to delete redundant and interfering points, further reducing data. Points are evaluated based on intensity, incidence angle, and recording distance. Depending on the object and point distance, data is reduced to 3–10% of the original size, making it more manageable and increasing its application range. For complex objects, segments can be manually created in the form of cuboids, prisms, or polyhedra. These segments allow selecting subsets of the point cloud for evaluation and can be organised to map the object’s structure (e.g., building, storey, room). Tools for automated or interactive determination of surface geometries (plane, cylinder, elliptical cylinder) were developed for spatial structure determination and modeling tasks. User coordinate systems (UCS) can be defined by selecting reference points, point sets, or horizontal intersection lines, forming the basis for CAD evaluations. Horizontally and vertically aligned sections can be generated. Initial approaches for automated vectorisation and dimensioning have been developed. Points can be orthogonally projected over planes or sections of developable surfaces for 2D visualisation. For deformation analysis, additional colour codes can be calculated relative to the distance from the reference plane, developed section, or reference point cloud. A VR interface was integrated for virtual inspection of scan data.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Proposed Solutions for the Automated Evaluation of Laser Scan Data

  • Gunnar Siedler,
  • Sebastian Vetter

摘要

When capturing an object using laser scanning, multiple viewpoints are utilised due to the recording perspective and required overlap for registration. During various R&D projects, methods were developed for the interactive and automated evaluation of merged point clouds in metigo® software. Pre-import of registered viewpoints allows loading all data in reduced quality to define a relevant object area, leading to significant initial data reduction. Individual viewpoints are merged into an overall point cloud with specified point spacing. A voxel grid, consisting of cube-shaped cells, is created to delete redundant and interfering points, further reducing data. Points are evaluated based on intensity, incidence angle, and recording distance. Depending on the object and point distance, data is reduced to 3–10% of the original size, making it more manageable and increasing its application range. For complex objects, segments can be manually created in the form of cuboids, prisms, or polyhedra. These segments allow selecting subsets of the point cloud for evaluation and can be organised to map the object’s structure (e.g., building, storey, room). Tools for automated or interactive determination of surface geometries (plane, cylinder, elliptical cylinder) were developed for spatial structure determination and modeling tasks. User coordinate systems (UCS) can be defined by selecting reference points, point sets, or horizontal intersection lines, forming the basis for CAD evaluations. Horizontally and vertically aligned sections can be generated. Initial approaches for automated vectorisation and dimensioning have been developed. Points can be orthogonally projected over planes or sections of developable surfaces for 2D visualisation. For deformation analysis, additional colour codes can be calculated relative to the distance from the reference plane, developed section, or reference point cloud. A VR interface was integrated for virtual inspection of scan data.