A Knowledge-driven Method for 3D Reconstruction of Technical Installations in Building Rehabilitations
Garmendia Aroztegui, Iñigo (2018)
2018
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2018-06-06
Julkaisun pysyvä osoite on
http://urn.fi/URN:NBN:fi:tty-201805241835
http://urn.fi/URN:NBN:fi:tty-201805241835
Tiivistelmä
Rehabilitation of buildings often requires analysing the scene and measuring the dimension of different elements. Normally this task is done by visiting the real scene. In many cases, the planes are lost or unavailable, so the area must be measured again. Building a 3D model of the scene enables to work remotely. The scene is captured using a 3D laser scanner, which creates a point cloud. Applying a segmentation method to the point cloud, the features of the data can be extracted.
In this thesis a method for processing the point cloud data obtaining the information of each wall and pipe is developed. The point cloud is cleaned of unnecessary points and segmented applying different segmentation methods. Region growing algorithm and RANSAC algorithms are combined for an accurate detection of pipes and planes. Once the features of each element are detected, a knowledge-based model, in this case an ontological model, is created. This ontological model provides information for building the 3D model in a web application. Furthermore, it contains data about the suppliers and products, which after applying some reasoning, the purchase options for a pipe is shown.
The research work is about scanning the scene with a 3D laser scanner and building an application for creating the model out of the scanned data. The web application recreates the structure and pipes on a 3D model, making the intersections between pipes and providing information about the replacement’s purchasing options.
In this thesis a method for processing the point cloud data obtaining the information of each wall and pipe is developed. The point cloud is cleaned of unnecessary points and segmented applying different segmentation methods. Region growing algorithm and RANSAC algorithms are combined for an accurate detection of pipes and planes. Once the features of each element are detected, a knowledge-based model, in this case an ontological model, is created. This ontological model provides information for building the 3D model in a web application. Furthermore, it contains data about the suppliers and products, which after applying some reasoning, the purchase options for a pipe is shown.
The research work is about scanning the scene with a 3D laser scanner and building an application for creating the model out of the scanned data. The web application recreates the structure and pipes on a 3D model, making the intersections between pipes and providing information about the replacement’s purchasing options.