Automatic reconstruction of 3D building models from scanned 2D floor plans

Gimenez, Lucile; Robert, Sylvain; Suard, Frédéric; Zreik, Khaldoun

doi:10.1016/j.autcon.2015.12.008

Cited by 97 publications

(73 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Ahmed et al [2] separated text from graphics and extracted lines of various thickness, where walls are extracted from the thicker lines and symbols are assumed to have thin lines; then, they applied such information to further locate doors and windows. Gimenez et al [7] recognized walls and openings using heuristics, and generated 3D building models based on the detected walls and doors.…”

Section: Related Workmentioning

confidence: 99%

Deep Floor Plan Recognition Using a Multi-Task Network With Room-Boundary-Guided Attention

Zeng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

101

View full text Add to dashboard Cite

This paper presents a new approach to recognize elements in floor plan layouts. Besides walls and rooms, we aim to recognize diverse floor plan elements, such as doors, windows and different types of rooms, in the floor layouts. To this end, we model a hierarchy of floor plan elements and design a deep multi-task neural network with two tasks: one to learn to predict room-boundary elements, and the other to predict rooms with types. More importantly, we formulate the room-boundary-guided attention mechanism in our spatial contextual module to carefully take room-boundary features into account to enhance the room-type predictions. Furthermore, we design a cross-and-within-task weighted loss to balance the multi-label tasks and prepare two new datasets for floor plan recognition. Experimental results demonstrate the superiority and effectiveness of our network over the state-of-the-art methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Deep Floor Plan Recognition Using a Multi-Task Network With Room-Boundary-Guided Attention

Zeng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

101

View full text Add to dashboard Cite

show abstract

“…[7] CVC-FP Qgar project, Hough transform Entity segmentation [9] CVC-FP Predefined rule Room detection [10] CVC-FP Contour extraction Room detection [11] Defined in paper Predefined rule Room detection [12] CVC-FP SVM-BoVM, centerline detection Room detection [13] CVC-FP, R-FP Deep learning (FCN-2s) Pixel [14] R-FP Deep learning (ResNet-152) Vector with points and lines Note: FCN-2s: Fully Convolutional Network with stride 2; ResNet-152: Residual Network 152.…”

Section: Dataset Methods Resultsmentioning

confidence: 99%

“…In addition, they could be difficult to use as spatial information due to their lack of coordinates. Gimenez et al [9] defined the category of wall segments well and classified them at the pixel level using a predefined rule set. However, as these methods have many predefined hyper-parameters, they need to be chosen before the process, and thus they lack generalizability.…”

Section: Introductionmentioning

confidence: 99%

Indoor Reconstruction from Floorplan Images with a Deep Learning Approach

Jang

Yang

2020

IJGI

View full text Add to dashboard Cite

Although interest in indoor space modeling is increasing, the quantity of indoor spatial data available is currently very scarce compared to its demand. Many studies have been carried out to acquire indoor spatial information from floorplan images because they are relatively cheap and easy to access. However, existing studies do not take international standards and usability into consideration, they consider only 2D geometry. This study aims to generate basic data that can be converted to indoor spatial information using IndoorGML (Indoor Geography Markup Language) thick wall model or the CityGML (City Geography Markup Language) level of detail 2 by creating vector-formed data while preserving wall thickness. To achieve this, recent Convolutional Neural Networks are used on floorplan images to detect wall and door pixels. Additionally, centerline and corner detection algorithms were applied to convert wall and door images into vector data. In this manner, we obtained high-quality raster segmentation results and reliable vector data with node-edge structure and thickness attributes that enabled the structures of vertical and horizontal wall segments and diagonal walls to be determined with precision. Some of the vector results were converted into CityGML and IndoorGML form and visualized, demonstrating the validity of our work.

show abstract

“…This is considered a supervised pattern recognition approach which is commonly used in Scan-vs-BIM (Bosché et al, 2013, Bosché et al, 2014. In this research, we propose a bottom-up method since the majority of built structures do not have reliable plan information (Gimenez et al, 2015). Furthermore, we solely rely on the point cloud and building logic for the point cloud interpretation since the incorporation of e.g.…”

Section: Introductionmentioning

confidence: 99%

Bim Reconstruction: Automated Procedural Modeling From Point Cloud Data

Bassier

Mattheuwsen

Vergauwen

2019

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

The reconstruction of Building Information Modeling objects for as-built modeling is currently the subject of ongoing research. A popular method is to extract structure information from point cloud data to create a set of parametric objects. This requires the interpretation of the point cloud data which currently is a manual and labor intensive procedure. Automated processes have to cope with excessive occlusions and clutter in the data sets. To create an as-built BIM, it is vital to reconstruct the building's structure i.e. wall geometry prior to the reconstruction of other objects. In this work, a novel method is presented to automatically reconstruct as-built BIM for generic buildings. We presented an unsupervised method that procedurally models the geometry of the walls based on point cloud data. A bottom-up process is defined where consecutively higher level information is extracted from the point cloud data using pre-trained machine learning models. Prior to the reconstruction, the data is segmented, classified and clustered to retrieve all the available observations of the walls. The resulting geometry is processed by the reconstruction algorithm. First, the necessary information is extracted from the observations for the creation of parametric solid objects. Subsequently, the final walls are created by updating their topology. The method is tested on a variety of scenes and shows promising results to reliably and accurately create as-built models. The accuracy of the generated geometry is similar to the precision of expert modelers. A key advantage is that that the algorithm creates Revit and Rhino native objects which makes the geometry directly applicable to a wide range of applications.This contribution has been peer-reviewed. https://doi.org/10.5194/isprs-archives-XLII-2-W17-53-2019 |

show abstract

Automatic reconstruction of 3D building models from scanned 2D floor plans

Cited by 97 publications

References 11 publications

Deep Floor Plan Recognition Using a Multi-Task Network With Room-Boundary-Guided Attention

Deep Floor Plan Recognition Using a Multi-Task Network With Room-Boundary-Guided Attention

Indoor Reconstruction from Floorplan Images with a Deep Learning Approach

Bim Reconstruction: Automated Procedural Modeling From Point Cloud Data

Contact Info

Product

Resources

About