Laser-based survey of building objects and buildings

Abstract

Building information models (BIMs) for facility management is gaining interest. Different technologies for collecting the raw material to extract such model are in rapid development. The most common technologies are based on images, structure light, laser or a combination of these. The new technologies have the potential to provide efficient data collection, but not necessarily at the same accuracy compared to the traditional methods. This thesis has explored how to rapidly establish a BIM for an existing building. This was done by investigating two different aspects related to this task. The first aspect was related to product specification and provide a framework for ordering and controlling a laser-based survey of a building. The second aspect explores how a laser-based system could be used to rapidly survey an existing building. Through the thesis and the first aspect, it is shown that the Norwegian survey community is lacking an unambiguous product specification for building surveys performed for BIM extraction and that the survey seldomly is adequately controlled. Based on these findings a product specification has been developed in cooperation with building owners. This cooperation made it possible to test the product specification in real projects. The product specification was developed through three different versions. The zero version was presented at the World Building Congress in 2016 and was tested in a renovation project at the Norwegian University of Life Sciences. The evaluation of the project led to the first version that was used in a framework competition arranged by Ullensaker municipality in the south-east of Norway. The result led to the second and final version of the product specification. The proposed product specification follows a simplified transaction pattern between the customer and the producer. The focus has been on the customer's request for a building survey suitable for BIM extraction and the customer's acceptance actions when the building survey is delivered. The acceptance actions are based on well–known standards created by the Norwegian Mapping Authority. The customer request is based on the acceptance actions. This ensures that every requirements can be verified in the accepting stage. The main purposes of the product specification were to ensure reliable results and to minimize the difference between the customer request and the producer’s delivery. Additionally, an unambiguous product specification can ensure a fair competition situation between the producers and give the producers the possibility to select the best-suited technology. The second aspect is related to how a building can be efficiently surveyed and explores how this could be done with a laser-based system. A human carried survey system was developed through three stages. The first and second stages focused on circle shaped objects and were realized in cooperation with the Faculty of Environmental Sciences and Natural Resource Management at the Norwegian University of Life Sciences. The system surveyed tree diameter at breast height within sample plots in size 250-500 m2. The system was able to detect 87.5% of the trees with a mean difference of 0.1 cm, and a root mean square of 2.2 cm. The novel aspect is related to how the trees are segmented and how the diameters are estimated without losing precision due to degraded pose solution. The result can be used in forestry inventory projects together with airborne laser surveys. The third stage was made for indoor measurements. The main focus was on how to aid the navigation solution in the absence of Global Navigation Satellite System signals. The method divides the laser point measurements into small time frames. For each time frame, the laser points were automatically classified into column, walls, floor, and ceiling. This information was used to support a scan matching method called semantic-assisted normal distributions transform. The result from the scan matching was used to create a trajectory of the walking path followed during data capture. This result was fed back into the inertial navigation processing to aid the solution when the system was located inside the building. This gives the inertial navigation process the ability to reject scan matching failures. The novel method was able to improve the survey accuracy from a maximum deviation of 12.6 m to 1.1 m. The third stage had two different Inertial Measurement Units (IMU) installed. The most accurate system was a tactical graded IMU, and the lowest accurate system was an automotive graded IMU. With the proposed method, the automotive graded system was able to perform at a higher level than a standalone tactical graded solution.

Interessen for å bruke BygningsInformasjonsModeller (BIMer) i forvaltning, drift og vedlikehold av bygninger er økende. Ulike teknologier for innsamling av data for å etablere slike modeller er i rask utvikling. De vanligste teknologiene er basert på bilder, strukturert lys, laser eller en kombinasjon av disse. Ny teknologi utfører målingene veldig effektivt, men ikke med samme nøyaktighet som tradisjoneller metoder. Denne studien har undersøkt hvordan en raskt kan etablere en BIM i et eksisterende bygg. Dette ble gjort ved å utforske to ulike aspekter av problemstillingen. Det første aspektet ser på produktspesifikasjon og foreslår et rammeverk til bruk ved bestilling og kontroll av laser-basert innmåling av eksisterende bygning. Det andre aspektet utforsker hvordan et laser-basert system raskt kan måle opp eksisterende bygg.