Reconstruction Of Building Models Research Articles

A fast point cloud registration method based on spatial relations and features

Abstract Point cloud registration plays a crucial role in mobile robot localization, map building and 3D model reconstruction. However, it remains challenged by issues such as compromised accuracy and sluggish efficiency, posing significant obstacles in achieving precise and timely alignments. Therefore, we propose a lightweight and fast point cloud registration method. Firstly, we mesh the 3D point cloud, compared with the traditional gridded point cloud method, it achieves initial point cloud registration by preserving the curvature characteristics of the internal point cloud, and utilizing the spatial relationship between grid cells and the quantitative relationship between the internal point cloud. Moreover, we adopt an iterative nearest point (ICP) based on KD-Tree to realize the fine registration. So, our method does not necessitate intricate feature analysis and data training, and is resilient to similar transformations, non-uniform densities and noise. Finally, we conduct point cloud registration experiments using multiple publicly available point cloud datasets and compare them with several point cloud registration methods. The results demonstrate it is able to accomplish the point cloud registration quickly and exhibit high accuracy. More importantly, it maintains its efficacy and robustness even in the presence of noisy and defective point clouds.

PLANES4LOD2: Reconstruction of LoD-2 building models using a depth attention-based fully convolutional neural network

Level of detail (LoD)-2 reconstruction is an inevitable task in digital twin-related applications such as disaster management, flood simulation, landslide simulation and solar panel recommendation. However, there is a lack of capable methods that can exploit fine details in RGB imagery and mitigate noise in photogrammetric digital surface models (DSMs). Our investigation is focused on the use of roof planes to achieve a geometrically complete and correct, and topologically consistent LoD-2 building reconstruction. Using UNet with the EfficientNet-B3 backbone, the developed approach starts with jointly predicting building sections and roof planes from the orthorectified RGB imagery and a photogrammetric DSM. The detected sections and planes are then vectorized by employing tree search and simplified with the Douglas Peucker algorithm. Subsequently, height values from the noisy input DSM and the vectorized image-based (and simplified) roof planes are used to derive 3D-planes. Finally, the building model is formed by computing plane intersections as the ridge lines. This study demonstrates that a well-designed depth attention module (DAM), which is the bottleneck of the UNet, can achieve a very good use of both spectral and depth features. The resultant 1-to-n correspondence between building section and roof plane benefits accurate and consistent building model reconstruction. Furthermore, it leads to a superior generalization capability of the proposed method. Experiments with 1437 buildings from the cities Cologne and Braunschweig, Germany, demonstrate the success of the proposed workflow in reconstructing compound buildings with complex roof structures. The achieved geometric mean absolute error (MAE) is 1.06m and 0.24m respectively. Comprehensive comparative evaluations showcase the superiority of the approach in terms of geometric completeness and accuracy, and topological consistence with. The improvement over SAT2LOD2 (Gui and Qin, 2021) is 1.12m in Cologne (data accessible at https://github.com/dlrPHS/GPUB) and 0.47m in Braunschweig in geometrical MAE.

Iterative Low-Poly Building Model Reconstruction from Mesh Soups Based on Contour

Existing contour-based building-reconstruction methods face the challenge of producing low-poly results. In this study, we introduce a novel iterative contour-based method to reconstruct low-poly meshes with only essential details from mesh soups. Our method focuses on two primary targets that determine the quality of the results: reduce the total number of contours, and generate compact surfaces between contours. Specifically, we implemented an iterative pipeline to gradually extract vital contours by loss and topological variance, and potential redundant contours will be removed in a post-processing procedure. Based on these vital contours, we extracted the planar primitives of buildings as references for contour refinement to obtain compact contours. The connection relationships between these contours are recovered for surface generation by a contour graph, which is constructed using multiple bipartite graphs. Then, a low-poly mesh can be generated from the contour graph using our contour-interpolation algorithm based on polyline splitting. The experiments demonstrated that our method produced satisfactory results and outperformed the previous methods.

Regularity-constrained point cloud reconstruction of building models via global alignment

Regularity-constrained point cloud reconstruction of building models via global alignment

Extraction of Building Roof Contours from Airborne LiDAR Point Clouds Based on Multidirectional Bands

Because of the complex structure and different shapes of building contours, the uneven density distribution of airborne LiDAR point clouds, and occlusion, existing building contour extraction algorithms are subject to such problems as poor robustness, difficulty with setting parameters, and low extraction efficiency. To solve these problems, a building contour extraction algorithm based on multidirectional bands was proposed in this study. Firstly, the point clouds were divided into bands with the same width in one direction, the points within each band were vertically projected on the central axis in the band, the two projection points with the farthest distance were determined, and their corresponding original points were regarded as the roof contour points; given that the contour points obtained based on single-direction bands were sparse and discontinuous, different banding directions were selected to repeat the above contour point marking process, and the contour points extracted from the different banding directions were integrated as the initial contour points. Then, the initial contour points were sorted and connected according to the principle of joining the nearest points in the forward direction, and the edges with lengths greater than a given threshold were recognized as long edges, which remained to be further densified. Finally, each long edge was densified by selecting the noninitial contour point closest to the midpoint of the long edge, and the densification process was repeated for the updated long edge. In the end, a building roof contour line with complete details and topological relationships was obtained. In this study, three point cloud datasets of representative building roofs were chosen for experiments. The results show that the proposed algorithm can extract high-quality outer contours from point clouds with various boundary structures, accompanied by strong robustness for point clouds differing in density and density change. Moreover, the proposed algorithm is characterized by easily setting parameters and high efficiency for extracting outer contours. Specific to the experimental data selected for this study, the PoLiS values in the outer contour extraction results were always smaller than 0.2 m, and the RAE values were smaller than 7%. Hence, the proposed algorithm can provide high-precision outer contour information on buildings for applications such as 3D building model reconstruction.

Shape-preserving mesh decimation for 3D building modeling

We propose a shape-preserving building model reconstruction method that involves simplifying the original building mesh to accommodate various building shapes and complexities. To achieve this, we apply a structure-aware segmentation technique to parse the ubiquitous building points into building geometric primitives and building structural points, i.e., anchor points. After that, we generate dense building meshes from building semantic points in a topology-aware manner. As the geometric primitive semantics are assigned to the building points during the structure-aware segmentation process, these primitive semantics of the building points can be explicitly transferred into the created building meshes. To offer lightweight and accurate building models with enriched semantics, we leverage the building structural points as constraints for the subsequent edge collapse simplification algorithm. This algorithm effectively decimates irrelevant vertices and meshes, while preserving the essential building structural contours. The entire simplification process is performed in a shape-preserving manner, granting us flexible control over the imposition of different degrees of strength regarding various geometric primitives during the simplification. We conduct qualitative and quantitative analyses to evaluate the effectiveness of our method on both individual building models and a large-scale urban scene. Additionally, we extensively compare our proposed shape-preserving algorithm with other state-of-the-art mesh decimation methods to demonstrate our superiority.

Optimal Pre-processing of Laser Scanning Data for Indoor Scene Analysis and 3D Reconstruction of Building Models

Optimal Pre-processing of Laser Scanning Data for Indoor Scene Analysis and 3D Reconstruction of Building Models

EVALUATING GEOMETRY OF AN INDOOR SCENARIO WITH OCCLUSIONS BASED ON TOTAL STATION MEASUREMENTS OF WALL ELEMENTS

Abstract. Scan2BIM approaches, i.e. the automated reconstruction of building models from point cloud data, is typically evaluated against the same point clouds which are used as input for the reconstruction process. In doing so, the point clouds are often used as ground truth without considering their own inaccuracies. Thus, in this research, we investigate the manual creation of an accurate ground truth, with a process which takes into account the measurement accuracy as well as the modeling accuracy. Therefore we created a ground truth to an existing laser scan data with a total station, based on the assumption that a total station generally measures points more reliably. In addition, a manual selection and classification of points on the wall surfaces during the measurement, serves a reliable detection of the walls via plane fitting. This allows for the creation of a more reliable ground truth, which is determined by the intersection of the planes from corners and edges. The ground truth is aligned parallel to the axes of a local coordinate system. From MLS and TLS point clouds of the same building area, walls are manually classified and corners and edges are determined in a similar way to the total station. These TLS and MLS corners are registered to this ground truth using least squares optimisation at the vertices. The transformation thus determined is used to transform the laser scanning point clouds as well. The resulting errors in the corners and the whole point cloud are evaluated. We conclude that the standard deviation of wall surfaces alone isn’t sufficient to determine the quality of the reconstructed building model. Despite low measurement noise in single wall surfaces, deviations in the reconstructed room model may arise.

Building element recognition with MTL-AINet considering view perspectives

Abstract The reconstruction and analysis of building models are crucial for the construction of smart cities. A refined building model can provide a reliable data support for data analysis and intelligent management of smart cities. The colors, textures, and geometric forms of building elements, such as building outlines, doors, windows, roof skylights, roof ridges, and advertisements, are diverse; therefore, it is challenging to accurately identify the various details of buildings. This article proposes the Multi-Task Learning AINet method that considers features such as color, texture, direction, and roll angle for building element recognition. The AINet is used as the basis function; the semantic projection map of color and texture, and direction and roll angle is used for multi-task learning, and the complex building facade is divided into similar semantic patches. Thereafter, the multi-semantic features are combined using hierarchical clustering with a region adjacency graph and the nearest neighbor graph to achieve an accurate recognition of building elements. The experimental results show that the proposed method has a higher accuracy for building detailed edges and can accurately extract detailed elements.

High-Precision Single Building Model Reconstruction Based on the Registration between OSM and DSM from Satellite Stereos

For large-scale 3D building reconstruction, there have been several approaches to utilizing multi-view satellite imagery to produce a digital surface model (DSM) for height information and extracting building footprints for contour information. However, limited by satellite resolutions and viewing angles, the corresponding DSM and building footprints are sometimes of a low accuracy, thus generating low-accuracy building models. Though some recent studies have added GIS data to refine the contour of the building footprints, the registration errors between the GIS data and satellite images are not considered. Since OpenStreetMap (OSM) provides a high level of precision and complete building polygons in most cities worldwide, this paper proposes an automatic single building reconstruction method that utilizes a DSM from high-resolution satellite stereos, as well as building footprints from OSM. The core algorithm accurately registers the building polygons from OSM with the rasterized height information from the DSM. To achieve this goal, this paper proposes a two-step “coarse-to-fine registration” algorithm, with both steps being formulated into the optimization of energy functions. The coarse registration is optimized by separately moving the OSM polygons at fixed steps with the constraints of a boundary gradient, an interior elevation mean, and variance. Given the initial solution of the coarse registration, the fine registration is optimized by a genetic algorithm to compute the accurate translations and rotations between the DSM and OSM. Experiments performed in the Beijing/Shanghai region show that the proposed method can significantly improve the IoU (intersection over union) of the registration results by 69.8%/26.2%, the precision by 41.0%/15.5%, the recall by 41.0%/16.0%, and the F1-score by 42.7%/15.8%. For the registration, the method can reduce the translation errors by 4.656 m/2.815 m, as well as the rotation errors by 0.538°/0.228°, which indicates its great potential in smart 3D applications.

Reconstruction of LoD-2 Building Models Guided by Façade Structures from Oblique Photogrammetric Point Cloud

Due to the façade visibility, intuitive expression, and multi-view redundancy, oblique photogrammetry can provide optional data for large-scale urban LoD-2 reconstruction. However, the inherent noise in oblique photogrammetric point cloud resulting from the image-dense matching limits further model reconstruction applications. Thus, this paper proposes a novel method for the efficient reconstruction of LoD-2 building models guided by façade structures from an oblique photogrammetric point cloud. First, a building planar layout is constructed combined with footprint data and the vertical planes of the building based on spatial consistency constraints. The cells in the planar layout represent roof structures with a distinct altitude difference. Then, we introduce regularity constraints and a binary integer programming model to abstract the façade with the best-fitting monotonic regularized profiles. Combined with the planar layout and regularized profiles, a 2D building topology is constructed. Finally, the vertices of building roof facets can be derived from the 2D building topology, thus generating a LoD-2 building model. Experimental results using real datasets indicate that the proposed method can generate reliable reconstruction results compared with two state-of-the-art methods.

3D RECONSTRUCTION OF SIMPLE BUILDINGS FROM POINT CLOUDS USING NEURAL NETWORKS WITH CONTINUOUS CONVOLUTIONS (CONVPOINT)

Abstract. The automatic reconstruction of 3D building models from airborne laser scanning point clouds or aerial imagery data in a model-driven fashion most often consists of a recognition of standardized building primitives with typically rectangular footprints and parameterized roof shapes based on a pre-defined collection, and a parameter estimation so that the selected primitives best fit the input data. For more complex buildings that consist of multiple parts, several such primitives need to be combined. This paper focuses on the reconstruction of such simple buildings, and explores the potential of Deep Learning by presenting a neural network architecture that takes a 3D point cloud of a single building as input and outputs the geometric information needed to construct a 3D building model in half-space representation with up to four roof faces like saddleback, hip, and pyramid roof. The proposed neural network architecture consists of a roof face segmentation module implemented with continuous convolutions as used in ConvPoint, which performs feature extraction directly from a set of 3D points, and four PointNet modules that predict from sampled subsets of the feature-enriched points the presence of four roof faces and their slopes. Trained with the RoofN3D dataset, which contains roof point segmentations and geometric information for 3D reconstruction purposes for about 118,000 simple buildings, the neural network achieves a performance of about 80% intersection over union (IoU) for roof face segmentation, 1.8° mean absolute error (MAE) for roof slope angles, and 95% overall accuracy (OA) for predicting the presence of faces.

3D building reconstruction from single street view images using deep learning

3D building models are an established instance of geospatial information in the built environment, but their acquisition remains complex and topical. Approaches to reconstruct 3D building models often require existing building information (e.g. their footprints) and data such as point clouds, which are scarce and laborious to acquire, limiting their expansion. In parallel, street view imagery (SVI) has been gaining currency, driven by the rapid expansion in coverage and advances in computer vision (CV), but it has not been used much for generating 3D city models. Traditional approaches that can use SVI for reconstruction require multiple images, while in practice, often only few street-level images provide an unobstructed view of a building. We develop the reconstruction of 3D building models from a single street view image using image-to-mesh reconstruction techniques modified from the CV domain. We regard three scenarios: (1) standalone single-view reconstruction; (2) reconstruction aided by a top view delineating the footprint; and (3) refinement of existing 3D models, i.e. we examine the use of SVI to enhance the level of detail of block (LoD1) models, which are common. The results suggest that trained models supporting (2) and (3) are able to reconstruct the overall geometry of a building, while the first scenario may derive the approximate mass of the building, useful to infer the urban form of cities. We evaluate the results by demonstrating their usefulness for volume estimation, with mean errors of less than 10% for the last two scenarios. As SVI is now available in most countries worldwide, including many regions that do not have existing footprint and/or 3D building data, our method can derive rapidly and cost-effectively the 3D urban form from SVI without requiring any existing building information. Obtaining 3D building models in regions that hitherto did not have any, may enable a number of 3D geospatial analyses locally for the first time.

RANSAC-based multi primitive building reconstruction from 3D point clouds

Building model reconstruction from 3D point clouds has been investigated for several decades with increasing interests. Building models represented by one or more parametric primitives can assure regularity and provide semantic information for the reconstructed buildings. However, there exist challenges in reliably determining building primitives, especially for compound buildings with multiple primitives. This paper presents a multi primitive reconstruction (MPR) approach to segment a compound bounding into several predefined primitives and determine their parameters from the point clouds. The method consists of primitive segmentation through a two-step RANSAC strategy, followed by holistic primitive fitting, and 3D Boolean operations. The first step segments the point cloud of a building into planar patches. The second step applies RANSAC strategy to further segment the predefined building primitives, where only points on adjacent planar patches are sampled to achieve high computational efficiency. The most probable primitive is then selected based on a set of quality metrics and corresponding parameters are holistically estimated with the identified inliers to form the building model. Finally, the 3D Boolean operation is used to reconstruct a topologically consistent 3D building model from its compositional primitives. The proposed RANSAC-MPR method has following advantages. (1) The framework for primitive segmentation is efficient since the sampling only occurs to adjacent planar patches; (2) the type of building primitives can be identified based on a score function without using advanced learning process; (3) compound buildings can be reconstructed through 3D union of the primitives determined by holistic fitting. Tested with 1054 buildings in three lidar and photogrammetry point clouds, the development is able to produce compound building models with regularized primitives at 85% boundary consistency and overall accuracy of 7 cm, which is about 0.14 times and 0.56 times ground point spacing for the lidar and photogrammetry datasets, respectively.

Pose Normalization of Indoor Mapping Datasets Partially Compliant with the Manhattan World Assumption

Due to their great potential for a variety of applications, digital building models are well established in all phases of building projects. Older stock buildings however frequently lack digital representations, and creating these manually is a tedious and time-consuming endeavor. For this reason, the automated reconstruction of building models from indoor mapping data has arisen as an active field of research. In this context, many approaches rely on simplifying suppositions about the structure of buildings to be reconstructed such as, e.g., the well-known Manhattan World assumption. This however not only presupposes that a given building structure itself is compliant with this assumption, but also that the respective indoor mapping dataset is aligned with the coordinate axes. Indoor mapping systems, on the other hand, typically initialize the coordinate system arbitrarily by the sensor pose at the beginning of the mapping process. Thus, indoor mapping data need to be transformed from the local coordinate system, resulting from the mapping process, to a local coordinate system where the coordinate axes are aligned with the Manhattan World structure of the building. This necessary preprocessing step for many indoor reconstruction approaches is also frequently known as pose normalization. In this paper, we present a novel pose-normalization method for indoor mapping point clouds and triangle meshes that is robust against large portions of the indoor mapping geometries deviating from an ideal Manhattan World structure. In the case of building structures that contain multiple Manhattan World systems, the dominant Manhattan World structure supported by the largest fraction of geometries was determined and used for alignment. In a first step, a vertical alignment orienting a chosen axis to be orthogonal to horizontal floor and ceiling surfaces was conducted. Subsequently, a rotation around the resulting vertical axis was determined that aligned the dataset horizontally with the axes of the local coordinate system. The performance of the proposed method was evaluated quantitatively on several publicly available indoor mapping datasets of different complexity. The achieved results clearly revealed that our method is able to consistently produce correct poses for the considered datasets for different input rotations with high accuracy. The implementation of our method along with the code for reproducing the evaluation is made available to the public.

Plane segmentation for a building roof combining deep learning and the RANSAC method from a 3D point cloud

The segmentation of a point cloud on the roof plane is of great significance to the reconstruction of building models. However, the traditional segmentation methods segment the aerial point cloud of the roof, which cannot fully express the geometric structure of the roof, whereas the deep learning-based methods have problems such as too much manual annotation and training time. In this work, a plane segmentation method for a building roof based on the PointNet network combined with the random sample consensus (RANSAC) algorithm is proposed to directly segment the whole point cloud of the building, but it is not limited to the point cloud of the roof. With the proposed framework, the roof part is extracted from the building by an improved PointNet network, and then the roof semantic point cloud is segmented by the RANSAC algorithm to complete the roof extraction. Based on the experimental results gained from multiple building point clouds, it is shown that the proposed method achieves the segmentation of a roof on most multi-plane roof building point clouds and that it has strong practical value.

QUALITY ASSESSMENT OF A NATIONWIDE DATA SET CONTAINING AUTOMATICALLY RECONSTRUCTED 3D BUILDING MODELS

Abstract. Fully automated reconstruction of high-detail building models on a national scale is challenging. It raises a set of problems that are seldom found when processing smaller areas, single cities. Often there is no reference, ground truth available to evaluate the quality of the reconstructed models. Therefore, only relative quality metrics are computed, comparing the models to the source data sets. In the paper we present a set of relative quality metrics that we use for assessing the quality of 3D building models, that were reconstructed in a fully automated process, in Levels of Detail 1.2, 1.3, 2.2 for the whole of the Netherlands. The source data sets for the reconstruction are the Dutch Building and Address Register (BAG) and the National Height Model (AHN). The quality assessment is done by comparing the building models to these two data sources. The work presented in this paper lays the foundation for future research on the quality control and management of automated building reconstruction. Additionally, it serves as an important step in our ongoing effort for a fully automated building reconstruction method of high-detail, high-quality models.

AUTOMATIC 3D BUILDING MODEL GENERATION USING DEEP LEARNING METHODS BASED ON CITYJSON AND 2D FLOOR PLANS

Abstract. In the past decade, a lot of effort is put into applying digital innovations to building life cycles. 3D Models have been proven to be efficient for decision making, scenario simulation and 3D data analysis during this life cycle. Creating such digital representation of a building can be a labour-intensive task, depending on the desired scale and level of detail (LOD). This research aims at creating a new automatic deep learning based method for building model reconstruction. It combines exterior and interior data sources: 1) 3D BAG, 2) archived floor plan images. To reconstruct 3D building models from the two data sources, an innovative combination of methods is proposed. In order to obtain the information needed from the floor plan images (walls, openings and labels), deep learning techniques have been used. In addition, post-processing techniques are introduced to transform the data in the required format. In order to fuse the extracted 2D data and the 3D exterior, a data fusion process is introduced. From the literature review, no prior research on automatic integration of CityGML/JSON and floor plan images has been found. Therefore, this method is a first approach to this data integration.

Combined Rule-Based and Hypothesis-Based Method for Building Model Reconstruction from Photogrammetric Point Clouds

Three-dimensional (3D) building models play an important role in digital cities and have numerous potential applications in environmental studies. In recent years, the photogrammetric point clouds obtained by aerial oblique images have become a major source of data for 3D building reconstruction. Aiming at reconstructing a 3D building model at Level of Detail (LoD) 2 and even LoD3 with preferred geometry accuracy and affordable computation expense, in this paper, we propose a novel method for the efficient reconstruction of building models from the photogrammetric point clouds which combines the rule-based and the hypothesis-based method using a two-stage topological recovery process. Given the point clouds of a single building, planar primitives and their corresponding boundaries are extracted and regularized to obtain abstracted building counters. In the first stage, we take advantage of the regularity and adjacency of the building counters to recover parts of the topological relationships between different primitives. Three constraints, namely pairwise constraint, triplet constraint, and nearby constraint, are utilized to form an initial reconstruction with candidate faces in ambiguous areas. In the second stage, the topologies in ambiguous areas are removed and reconstructed by solving an integer linear optimization problem based on the initial constraints while considering data fitting degree. Experiments using real datasets reveal that compared with state-of-the-art methods, the proposed method can efficiently reconstruct 3D building models in seconds with the geometry accuracy in decimeter level.

A new approach for roof segmentation from airborne LiDAR point clouds

ABSTRACT Accurate roof segmentation is one of the key steps for automatically constructing three-dimensional (3-D) building models. Building roofs can differ significantly in terms of their size, shape complexity, and number, rendering many existing airborne Light Detection And Ranging (LiDAR) roof segmentation methods ineffective. Thus, the applicability and precision of these methods need to be improved. For this purpose, this paper proposes a new roof segmentation method for airborne LiDAR point clouds. The proposed method integrates a novel region growing strategy and RANdom SAmple Consensus (RANSAC), applying these approaches to extract several reliable roof patches and then performing an iterative process to merge roof patches based on their parameters and the concept of inlier selection of RANSAC. Finally, unsegmented points and segmentation results are refined by voting in a local neighbourhood. The experimental results show that the proposed method can effectively segment the roofs of buildings with different complexity and sizes. As the basic evaluation primitive, the average segmentation correctness was found to be 95.67 and 97.85% when using a roof and a single point, respectively, which can provide reliable information for applications, such as 3-D building model reconstruction.