Surveyors, designers, construction crews and engineers all use point clouds. The possibilities that point clouds offer are only growing as technology advances. Anyone who works with 3D models can benefit from using point cloud data sets, and understanding best practices allows us to go even further.
A detailed and precise 3D model is based on several scans of the same area or the same object which must be aligned – a process called “assembly”. Assembly is a problematic process, as well as solutions of varying quality. This article provides an overview of some of the questions posed by creating point clouds and suggests a recent solution to some persistent problems.
“Status quo” point cloud
Scanners perform line of sight measurements. The amount of coverage required for a detailed 3D model usually requires multiple scans. These scans must be assembled to create a complete image of the area. Assembling the scans with each other poses the major question of creating point clouds.
In some cases, GNSS (Global Navigation Satellite System) technology such as GPS can be used to align scans using position data from scanners. But a given number of situations makes it impossible to record by GPS, for example by surveying the interior of an area, in narrow urban environments or in a forest.
In these cases, the computer software is able to analyze the data in the point cloud to find common geometries on which to assemble the scans. Two assembly methods were used, one using targets and the other relying solely on the use of software.
Targets are point-reducible artificial objects that are easily recognized by point cloud software for assembly. At least three targets must be placed for each pair of scans. The location of the targets must be planned, then the targets must be placed manually in the environment to be surveyed. Different environments present a variety of barriers to target placement, which requires time and the mobilization of qualified personnel.
In order to assemble the scans without targets, the geometries present in the digitized environment are used. This can save time in the field, but can add work time to the office.
Using targetless assembly software takes a long time to process many scans. On a 3D laser survey project, it takes a lot more time for the software to examine the point clouds for similarities to identify artificial targets. The considerable amount of data involved leads to longer processing, and each pair of scans must be evaluated according to three alignments: rotational, vertical and horizontal.
Standard targetless assembly software also requires operators to manually enter and verify each assembly, as distortions or inaccuracies in a single scan can impact the accuracy of the entire project. The need for manual intervention means more work and therefore more time.
Traditional targetless assembly requires up to 60% overlap to ensure precise assembly, which means targetless assembly requires more scans than target assembly.
Anyone who creates point cloud data sets faces the conundrum with or without targets, as both are time consuming and labor intensive. At ATIS and CYDIS, we have no assembly problem. Thanks to its SLAM technology, the PX-80 automatically assembles the point clouds it generates. To be able to fully exploit the results of the PX-80 mobile laser scanner, a SLAM calculation phase equivalent to 4 times the reading time must be undertaken. Thus for a 4,000 m² building, it will take 15 minutes to read it, ie 1 hour of SLAM calculation. The building’s point cloud will therefore be available within 1 hour and 15 minutes including reading. Add a few minutes for its transfer to the ATIS.cloud web platform and your point cloud will be accessible by all authorized persons.