Point cloud analysis for construction supervision

Challenge

Laser scanning is more often used for construction maintenance: a 3D laser scanner creates a point cloud of the structure, which is then compared with the original 3D model. This allows to detect deviations from the project plan on time and control contractors' work.

Comparing a point cloud with a model can take up to a week of working time since engineers have to manually identify every element in the cloud and outline all points which belong to it. The quality and speed of comparison directly depend on the engineer's qualification, while any mistake can affect the construction schedule and costs, or even lead to an accident on the construction site.

Gazprom Neft reached out to us with a task to develop a solution for automatic comparison of point clouds and 3D models, and fast identification of project deviations.

Rubius has developed a system that automatically compares a point cloud with a 3D model and finds inconsistencies in 30 seconds.

Cloud preprocessing — highlighting the elements belonging to the model, noise removal and downsampling to specified density

Automatic alignment of the point cloud with the model by 4 support points. The system selects these points at the intersection of massive horizontal and vertical planes (such as the ground or building walls) or according to a predetermined structural element

Cloud segmentation — identification of structural elements (beams, pipes, channels, wires, etc) and their classification according to the library of standard elements

Automatic comparison of a point cloud against a 3D model and detection of inconsistencies

Assessment of deviations by specified parameters (dimensions, spatial position, surface formation accuracy)

Generation of a report table with calculations of project deviations

Solution

Support of STL, FBX and PLY formats for 3D models, and PCD, PLY and LAS for point clouds

The system has the following functionality

Apart from construction supervision, the system can be used to examine structures or assess damage caused by accidents and emergencies.