scholarly journals Historic Timber Roof Structure Reconstruction through Automated Analysis of Point Clouds

2022 ◽  
Vol 8 (1) ◽  
pp. 10
Author(s):  
Taşkın Özkan ◽  
Norbert Pfeifer ◽  
Gudrun Styhler-Aydın ◽  
Georg Hochreiner ◽  
Ulrike Herbig ◽  
...  
Keyword(s):  
Laser Scanning ◽  
3D Model ◽  
Region Growing ◽  
Point Clouds ◽  
Automated Analysis ◽  
Structural Elements ◽  
Roof Structure ◽  
Main Target ◽  
Side Face ◽  
Region Growing Segmentation

We present a set of methods to improve the automation of the parametric 3D modeling of historic roof structures using terrestrial laser scanning (TLS) point clouds. The final product of the TLS point clouds consist of 3D representation of all objects, which were visible during the scanning, including structural elements, wooden walking ways and rails, roof cover and the ground; thus, a new method was applied to detect and exclude the roof cover points. On the interior roof points, a region-growing segmentation-based beam side face searching approach was extended with an additional method that splits complex segments into linear sub-segments. The presented workflow was conducted on an entire historic roof structure. The main target is to increase the automation of the modeling in the context of completeness. The number of manually counted beams served as reference to define a completeness ratio for results of automatically modeling beams. The analysis shows that this approach could increase the quantitative completeness of the full automatically generated 3D model of the roof structure from 29% to 63%.

scholarly journals AUTOMATED RECONSTRUCTION OF HISTORIC ROOF STRUCTURES FROM POINT CLOUDS – DEVELOPMENT AND EXAMPLES

2017 ◽  
Vol IV-2/W2 ◽  
pp. 195-202 ◽  
Author(s):  
M. Pöchtrager ◽  
G. Styhler-Aydın ◽  
M. Döring-Williams ◽  
N. Pfeifer
Keyword(s):  
Laser Scanning ◽  
Large Scale ◽  
Structural Model ◽  
Point Clouds ◽  
Automated Analysis ◽  
Structural Elements ◽  
Large Scale Structures ◽  
Scan Data ◽  
Short Time ◽  
High Degree

The analysis of historic roof constructions is an important task for planning the adaptive reuse of buildings or for maintenance and restoration issues. Current approaches to modeling roof constructions consist of several consecutive operations that need to be done manually or using semi-automatic routines. To increase efficiency and allow the focus to be on analysis rather than on data processing, a set of methods was developed for the fully automated analysis of the roof constructions, including integration of architectural and structural modeling. Terrestrial laser scanning permits high-detail surveying of large-scale structures within a short time. Whereas 3-D laser scan data consist of millions of single points on the object surface, we need a geometric description of structural elements in order to obtain a structural model consisting of beam axis and connections. Preliminary results showed that the developed methods work well for beams in flawless condition with a quadratic cross section and no bending. Deformations or damages such as cracks and cuts on the wooden beams can lead to incomplete representations in the model. Overall, a high degree of automation was achieved.

scholarly journals Segmentation and Multi-Scale Convolutional Neural Network-Based Classification of Airborne Laser Scanner Data

Sensors ◽  
2018 ◽  
Vol 18 (10) ◽  
pp. 3347 ◽  
Author(s):  
Zhishuang Yang ◽  
Bo Tan ◽  
Huikun Pei ◽  
Wanshou Jiang
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Laser Scanning ◽  
Laser Scanner ◽  
Region Growing ◽  
Point Clouds ◽  
Classification Method ◽  
Multi Scale ◽  
Airborne Laser

The classification of point clouds is a basic task in airborne laser scanning (ALS) point cloud processing. It is quite a challenge when facing complex observed scenes and irregular point distributions. In order to reduce the computational burden of the point-based classification method and improve the classification accuracy, we present a segmentation and multi-scale convolutional neural network-based classification method. Firstly, a three-step region-growing segmentation method was proposed to reduce both under-segmentation and over-segmentation. Then, a feature image generation method was used to transform the 3D neighborhood features of a point into a 2D image. Finally, feature images were treated as the input of a multi-scale convolutional neural network for training and testing tasks. In order to obtain performance comparisons with existing approaches, we evaluated our framework using the International Society for Photogrammetry and Remote Sensing Working Groups II/4 (ISPRS WG II/4) 3D labeling benchmark tests. The experiment result, which achieved 84.9% overall accuracy and 69.2% of average F1 scores, has a satisfactory performance over all participating approaches analyzed.

scholarly journals FAST AND ROBUST SEGMENTATION AND CLASSIFICATION FOR CHANGE DETECTION IN URBAN POINT CLOUDS

2016 ◽  
Vol XLI-B3 ◽  
pp. 693-699 ◽  
Author(s):  
X. Roynard ◽  
J.-E. Deschaud ◽  
F. Goulette
Keyword(s):  
Change Detection ◽  
Laser Scanning ◽  
Three Dimensional ◽  
Region Growing ◽  
Point Clouds ◽  
Object Class ◽  
Robust Segmentation ◽  
Street Furniture ◽  
Fast Region

Change detection is an important issue in city monitoring to analyse street furniture, road works, car parking, etc. For example, parking surveys are needed but are currently a laborious task involving sending operators in the streets to identify the changes in car locations. In this paper, we propose a method that performs a fast and robust segmentation and classification of urban point clouds, that can be used for change detection. We apply this method to detect the cars, as a particular object class, in order to perform parking surveys automatically. A recently proposed method already addresses the need for fast segmentation and classification of urban point clouds, using elevation images. The interest to work on images is that processing is much faster, proven and robust. However there may be a loss of information in complex 3D cases: for example when objects are one above the other, typically a car under a tree or a pedestrian under a balcony. In this paper we propose a method that retain the three-dimensional information while preserving fast computation times and improving segmentation and classification accuracy. It is based on fast region-growing using an octree, for the segmentation, and specific descriptors with Random-Forest for the classification. Experiments have been performed on large urban point clouds acquired by Mobile Laser Scanning. They show that the method is as fast as the state of the art, and that it gives more robust results in the complex 3D cases.

scholarly journals RGB-D Indoor Plane-based 3D-Modeling using Autonomous Robot

2014 ◽  
Vol XL-1 ◽  
pp. 301-308 ◽  
Author(s):  
N. Mostofi ◽  
A. Moussa ◽  
M. Elhabiby ◽  
N. El-Sheimy
Keyword(s):  
3D Modeling ◽  
3D Model ◽  
Research Work ◽  
Region Growing ◽  
Point Clouds ◽  
Visual Odometry ◽  
3D Models ◽  
Indoor Environments ◽  
Relative Pose ◽  
Matching Techniques

3D model of indoor environments provide rich information that can facilitate the disambiguation of different places and increases the familiarization process to any indoor environment for the remote users. In this research work, we describe a system for visual odometry and 3D modeling using information from RGB-D sensor (Camera). The visual odometry method estimates the relative pose of the consecutive RGB-D frames through feature extraction and matching techniques. The pose estimated by visual odometry algorithm is then refined with iterative closest point (ICP) method. The switching technique between ICP and visual odometry in case of no visible features suppresses inconsistency in the final developed map. Finally, we add the loop closure to remove the deviation between first and last frames. In order to have a semantic meaning out of 3D models, the planar patches are segmented from RGB-D point clouds data using region growing technique followed by convex hull method to assign boundaries to the extracted patches. In order to build a final semantic 3D model, the segmented patches are merged using relative pose information obtained from the first step.

scholarly journals INTEGRATION OF POINT CLOUDS DATASET FROM DIFFERENT SENSORS

2017 ◽  
Vol XLII-2/W3 ◽  
pp. 9-15 ◽  
Author(s):  
C. K. A. F. Che Ku Abdullah ◽  
N. Z. S. Baharuddin ◽  
M. F. M. Ariff ◽  
Z. Majid ◽  
C. L. Lau ◽  
...  
Keyword(s):  
Data Acquisition ◽  
Laser Scanning ◽  
3D Model ◽  
Accuracy Assessment ◽  
Three Dimensional ◽  
Laser Scanner ◽  
Point Clouds ◽  
Accurate Information ◽  
Integration Process ◽  
Building Facade

Laser Scanner technology become an option in the process of collecting data nowadays. It is composed of Airborne Laser Scanner (ALS) and Terrestrial Laser Scanner (TLS). ALS like Phoenix AL3-32 can provide accurate information from the viewpoint of rooftop while TLS as Leica C10 can provide complete data for building facade. However if both are integrated, it is able to produce more accurate data. The focus of this study is to integrate both types of data acquisition of ALS and TLS and determine the accuracy of the data obtained. The final results acquired will be used to generate models of three-dimensional (3D) buildings. The scope of this study is focusing on data acquisition of UTM Eco-home through laser scanning methods such as ALS which scanning on the roof and the TLS which scanning on building façade. Both device is used to ensure that no part of the building that are not scanned. In data integration process, both are registered by the selected points among the manmade features which are clearly visible in Cyclone 7.3 software. The accuracy of integrated data is determined based on the accuracy assessment which is carried out using man-made registration methods. The result of integration process can achieve below 0.04m. This integrated data then are used to generate a 3D model of UTM Eco-home building using SketchUp software. In conclusion, the combination of the data acquisition integration between ALS and TLS would produce the accurate integrated data and able to use for generate a 3D model of UTM eco-home. For visualization purposes, the 3D building model which generated is prepared in Level of Detail 3 (LOD3) which recommended by City Geographic Mark-Up Language (CityGML).

scholarly journals Change Detection of Urban Trees in MLS Point Clouds Using Occupancy Grids

2021 ◽  
Author(s):  
Philipp-Roman Hirt ◽  
Yusheng Xu ◽  
Ludwig Hoegner ◽  
Uwe Stilla
Keyword(s):  
Laser Scanning ◽  
Dimensional Space ◽  
Three Dimensional ◽  
Point Clouds ◽  
Automated Analysis ◽  
Urban Trees ◽  
Change Analysis ◽  
Single Tree ◽  
Local Point ◽  
Occupancy Grids

AbstractTrees play an important role in the complex system of urban environments. Their benefits to environment and health are manifold. Yet, especially near streets, the traffic can be impaired by a limited clearance. Even injuries could be caused by breaking tree parts. Hence, it is important to capture the trees in the frame of a tree cadastre and to ensure regular monitoring. Mobile laser scanning (MLS) can be used for data acquisition, followed by an automated analysis of the point clouds acquired over time. The presented approach uses occupancy grids with a grid size of 10 cm, which enable the comparison of several epochs in three-dimensional space. Prior to that, a segmentation of single tree objects is conducted. After cylinder-based trunk localisation, closely neighboured tree crowns are separated using weights derived from local point densities. Therefore, changes for every single tree can be derived with regard to its parameters and its point cloud. The testing area is set along an urban street in Munich, Germany, using the publicly available benchmark data sets TUM-MLS-2016/2018. In the frame of the evaluation, tree objects are geo-referenced and mapped in 2D. The tree parameters height and diameter at breast height are derived. The geometric evaluation of the change analysis facilitates not only the acquisition of stock changes, but also the detection of shape changes for the tree objects.

scholarly journals FAST AND ROBUST SEGMENTATION AND CLASSIFICATION FOR CHANGE DETECTION IN URBAN POINT CLOUDS

2016 ◽  
Vol XLI-B3 ◽  
pp. 693-699 ◽  
Author(s):  
X. Roynard ◽  
J.-E. Deschaud ◽  
F. Goulette
Keyword(s):  
Change Detection ◽  
Laser Scanning ◽  
Three Dimensional ◽  
Region Growing ◽  
Point Clouds ◽  
Object Class ◽  
Robust Segmentation ◽  
Street Furniture ◽  
Fast Region

Change detection is an important issue in city monitoring to analyse street furniture, road works, car parking, etc. For example, parking surveys are needed but are currently a laborious task involving sending operators in the streets to identify the changes in car locations. In this paper, we propose a method that performs a fast and robust segmentation and classification of urban point clouds, that can be used for change detection. We apply this method to detect the cars, as a particular object class, in order to perform parking surveys automatically. A recently proposed method already addresses the need for fast segmentation and classification of urban point clouds, using elevation images. The interest to work on images is that processing is much faster, proven and robust. However there may be a loss of information in complex 3D cases: for example when objects are one above the other, typically a car under a tree or a pedestrian under a balcony. In this paper we propose a method that retain the three-dimensional information while preserving fast computation times and improving segmentation and classification accuracy. It is based on fast region-growing using an octree, for the segmentation, and specific descriptors with Random-Forest for the classification. Experiments have been performed on large urban point clouds acquired by Mobile Laser Scanning. They show that the method is as fast as the state of the art, and that it gives more robust results in the complex 3D cases.

scholarly journals POINT CLOUD DERIVED FROMVIDEO FRAMES: ACCURACY ASSESSMENT IN RELATION TO TERRESTRIAL LASER SCANNINGAND DIGITAL CAMERA DATA

2017 ◽  
Vol XLII-2/W3 ◽  
pp. 217-223 ◽  
Author(s):  
P. Delis ◽  
M. Zacharek ◽  
D. Wierzbicki ◽  
A. Grochala
Keyword(s):  
Data Storage ◽  
Point Cloud ◽  
Laser Scanning ◽  
3D Model ◽  
Accuracy Assessment ◽  
Point Clouds ◽  
Histogram Equalization ◽  
Video Data ◽  
Video Frame ◽  
Video Frames

The use of image sequences in the form of video frames recorded on data storage is very useful in especially when working with large and complex structures. Two cameras were used in this study: Sony NEX-5N (for the test object) and Sony NEX-VG10 E (for the historic building). In both cases, a Sony α f = 16 mm fixed focus wide-angle lens was used. Single frames with sufficient overlap were selected from the video sequence using an equation for automatic frame selection. In order to improve the quality of the generated point clouds, each video frame underwent histogram equalization and image sharpening. Point clouds were generated from the video frames using the SGM-like image matching algorithm. The accuracy assessment was based on two reference point clouds: the first from terrestrial laser scanning and the second generated based on images acquired using a high resolution camera, the NIKON D800. The performed research has shown, that highest accuracies are obtained for point clouds generated from video frames, for which a high pass filtration and histogram equalization had been performed. Studies have shown that to obtain a point cloud density comparable to TLS, an overlap between subsequent video frames must be 85 % or more. Based on the point cloud generated from video data, a parametric 3D model can be generated. This type of the 3D model can be used in HBIM construction.

scholarly journals Comparison of Airborne Laser Scanning of Low and High Above Ground Level for Selected Infrastructure Objects

2018 ◽  
Vol 8 (2) ◽  
pp. 89-96 ◽  
Author(s):  
J. Siwiec
Keyword(s):  
Laser Scanning ◽  
New Technology ◽  
Ground Level ◽  
Point Clouds ◽  
Airborne Lidar ◽  
Structural Elements ◽  
Transmission Tower ◽  
Deviation Analysis ◽  
Airborne Laser ◽  
Internal Coherence

Abstract Along with the development of the technology of drone construction (UAV - Unmanned Aerial Vehicles), the number of applications of these solutions in the industry also grew. The aim of the research is to check the accuracy of data obtained using the new technology of UAV scanning and to compare them with one that is widely spread - high-altitude airborne Lidar, in terms of quality and spectrum of applications in industry and infrastructure. The research involved two infrastructure objects: a reinforced concrete one-span bridge and Lattice transmission tower with powerlines. The density of measurement, internal and external cohesion of point clouds obtained from both methods were compared. Plane fitting and deviation analysis were used. The data of UAV origin in both cases provided a sufficient density, allowing the recognition of structural elements, and internal coherence and precision of measurements important in modeling. The study shows that UAV mounted scanning may be used in the same applications as Airborne Lidar, as well as in other tasks requiring greater precision.

Sign in / Sign up

Export Citation Format

Share Document