Real Scan Data Research Articles

Failure mode and load prediction of steel bridge girders through 3D laser scanning and machine learning methods

AbstractCorrosion poses a significant threat to the longevity of steel bridges, impacting overall structural integrity. To effectively assess the structural condition of corroded steel bridges, conventional methods rely on visual inspections or single point measurements. To enhance and modernize this approach, this study introduces a novel framework integrating laser scanning data, computational models, and convolutional neural networks (CNNs). The CNN models are trained on a data set consisting of more than 1400 artificial corrosion scenarios generated by parameterizing real scan data from naturally corroded girders. This innovative method predicts the residual capacity and failure mode of corroded beam ends, achieving a low error rate of up to 3.3%. Unlike established evaluation procedures, the proposed evaluation framework directly utilizes post‐processed laser scanner output, eliminating the need for feature extraction and calculations.

As-built BIM reconstruction of piping systems using smartphone videogrammetry and terrestrial laser scanning

The increasing adoption of Building Information Modeling (BIM) in the construction industry has created a demand for methods to update existing facilities with accurate as-built BIM models. While terrestrial laser scanners (TLS) are commonly used for indoor surveys due to their speed and precision, they face limitations in areas with physical constraints like confined spaces and occlusions, with pipes being particularly affected. This paper presents a smartphone videogrammetry approach designed to enhance the accuracy and completeness of terrestrial laser scanner (TLS) point clouds for as-built BIM reconstruction. The method involves capturing missing areas within the TLS point cloud with smartphones and then seamlessly integrating these data into a more comprehensive multi-modal point cloud. This approach is validated using real scan data and offers convenience through the use of readily available and cost-effective smartphones, while its iterative nature facilitates progressive BIM reconstruction of pipe structures. Additionally, a compatible progressive BIM reconstruction framework is outlined, aiming to modernize scan-to-BIM practices in the construction industry, with a case study demonstrating its feasibility.

Mesh Total Generalized Variation for Denoising.

Recent studies have shown that the Total Generalized Variation (TGV) is highly effective in preserving sharp features as well as smooth transition variations for image processing tasks. However, currently there is no existing work that is suitable for applying TGV to 3D data, in particular, triangular meshes. In this article, we develop a novel framework for discretizing second-order TGV on triangular meshes. Further, we propose a TGV-based variational method for the denoising of face normal fields on triangular meshes. The TGV regularizer in our method is composed of a first-order term and a second-order term, which are automatically balanced. The first-order term allows our TGV regularizer to locate and preserve sharp features, while the second-order term allows our regularizer to recognize and recover smoothly curved regions. To solve the optimization problem, we introduce an efficient iterative algorithm based on variable-splitting and augmented Lagrangian method. Extensive results and comparisons on synthetic and real scanning data validate that the proposed method outperforms the state-of-the-art visually and numerically.

Object verification based on deep learning point feature comparison for scan-to-BIM

Building information models (BIMs) have been widely adopted in current construction projects to enhance the efficiency of facility maintenance operations. As-built BIMs can reflect the actual conditions of facilities and thus as-built BIM reconstruction has shown great significance in digital twin generation, building health monitoring, facility management and urban renewal. Laser scanners are capable to capture dense 3D measurements of the environment in a fast and highly accurate way. Therefore, laser scanning data have been widely used for as-built BIM generation. Although research efforts have been made on how to automatically achieve “Scan-to-BIM”, there are still gaps from applying current solutions to real scenarios. One of the challenges is that some irrelevant point clusters may be wrongly recognized as the desired object in the detection stage. This study presents a novel object verification approach based on deep learning point feature comparison to improve the accuracy of automated BIM reconstruction process. Firstly, a KPConv-based deep neural network is developed and trained to perform 3D point feature computation. Then through comparing point features calculated for extracted point clusters and as-designed BIM generated point clouds, point feature distance maps are generated. Afterwards, to automatically analyze the generated feature distance maps, a dataset including simulated positive and negative instances is created based on ModelNet40. And a tiny neural network is established and trained on the prepared dataset to acquire ability of distinguishment. To validate the feasibility of the proposed technique, experiments were conducted on both artificial point clouds and real scan data collected in one MEP room in a water treatment work in Hong Kong. It is demonstrated that the proposed technique can successfully filter out all the false positives in the Scan-to-BIM process, improving reconstruction accuracy significantly.

Automated geometry measurement and deep rolling of butt welds

During the joining of two metal sheets by welding, a process-specific geometry of the weld is created. The local geometry of the created weld has a decisive influence on its fatigue strength. This is due to stress concentration at the geometric notches. In this paper, a process known from mechanical engineering called deep rolling is applied on butt welds. The influence on the local weld geometry and the local stress concentration after deep rolling is investigated. Additionally, a novel automated measurement system using optical laser line scanning is presented. The system is qualified for the evaluation of the local weld geometry regarding its flank angles and toe radii. The presented investigations show that the deep rolling process influences the stress concentrations determined by 2D-FE-simulations using real scan data. A correlation between the difference in toe radii or local notch stresses before and after deep rolling and the initial flank angle was found. This indicates that there are process and geometry specific conditions for the successful application of the deep rolling process.

Multi-Scale Upsampling GAN Based Hole-Filling Framework for High-Quality 3D Cultural Heritage Artifacts

With the rapid development of 3D scanners, the cultural heritage artifacts can be stored as a point cloud and displayed through the Internet. However, due to natural and human factors, many cultural relics had some surface damage when excavated. As a result, the holes caused by these damages still exist in the generated point cloud model. This work proposes a multi-scale upsampling GAN (MU-GAN) based framework for completing these holes. Firstly, a 3D mesh model based on the original point cloud is reconstructed, and the method of detecting holes is presented. Secondly, the point cloud patch contains hole regions and is extracted from the point cloud. Then the patch is input into the MU-GAN to generate a high-quality dense point cloud. Finally, the empty areas on the original point cloud are filled with the generated dense point cloud patches. A series of real-world experiments are conducted on real scan data to demonstrate that the proposed framework can fill the holes of 3D heritage models with grained details. We hope that our work can provide a useful tool for cultural heritage protection.

Tiny hole inspection of aircraft engine nacelle in 3D point cloud via robust statistical fitting

In the aircraft manufacturing industry, drilling hole inspection is a vital task for the noise reduction ability and the aircraft structure stability. Due to the complexity of the inner surface of the engine nacelle, inspecting drilling holes of a small size and a large number is fairly challenging. In this paper, we propose a framework for automatic hole inspection on composite flat parts. The raw data of traditional 3D laser scanning usually contain considerable noise and outliers, which has a great influence on tiny hole inspection. First, to perform measurement efficiently and to get raw data of good quality, we design a measurement platform to perform automatic measurement on a preset path. Instead of applying common boundary detection methods, we design a method to detect boundary points by evaluating the boundary possibility of each point, which is more likely to detect the points on a real hole boundary than noise. Based on the characteristics of laser scanning data, we then combine an effective algebraic circle fitting method with an iterative fitting strategy and develop a robust circle fitting method. The improved circle fitting method is capable of estimating a circle closest to the ground truth. Experiments demonstrate that our fitting method performs better than other common fitting methods for comparison under both synthetic and real scanning data.

Point Cloud Upsampling via Perturbation Learning

Given sparse point clouds, this paper develops a perturbation learning-based point cloud upsampling method to generate uniform, clean, and dense point clouds. We build a simple yet efficient neural network framework including feature extraction, perturbation learning, and coordinate reconstruction operations. In the feature extraction task, shallow-and-wide dense connections are applied to present the latent geometric information. Subsequently, the extracted features are expanded for perturbation learning. According to the theory of the differential geometry of surfaces, the position of an upsampled point can be approximated by its projection on a tangent plane in a sufficiently small neighborhood around the point. Inspired by this, we propose learning a 2D perturbation through multilayer perceptrons (MLPs) to estimate the coordinate shift from the input point to the upsampled point. Then, we concatenate the 2D perturbation with the extracted features for residual learning to fine-tune the coordinate shift. Finally, the coordinate reconstruction step transforms all the high-level features into an upsampled and consolidated point cloud. To enable the learning-based point cloud upsampling process above, we collect a large-scale point cloud dataset that contains 36000 pairs of training sets. The entire network size is only 5.01 MB, which is much smaller than the requirements of state-of-the-art point cloud upsampling models. Qualitative and quantitative evaluation results show that our proposed scheme outperforms most existing methods in terms of the Chamfer distance (CD), Hausdorff distance (HD), Jensen-Shannon divergence (JSD), and uniformity. In addition, our method is applied to real scan data, and its robustness is confirmed.

Geometrical model based scan planning approach for the classification of rebar diameters

The inspection of rebar diameter prior to concrete pouring is a crucial step for inspectors and fabricators to check the compliance to the designed bearing capacity and structural integrity of reinforced concrete structures during the stages of manufacturing and construction. Current practices for dimensional rebar inspection are, however, conducted manually using measurement tapes, which is labor-intensive and error prone. The authors' group have previously proposed laser scanning-based methods to perform rebar diameter and spacing estimation using laser scanning technologies. However, those prior studies have limitations in not only accurately classifying small-size rebar diameters but also determining optimal locations of laser scanner to ensure accurate rebar dimensional inspection. This study presents a scan planning method to select the optimal scan position of laser scanner for enhancing the rebar inspection accuracy. To this end, a model which defines the geometrical relationship between the rebar layout and the laser scanner is first developed to simulate the coordinates of scan points falling onto the rebar surface. Scan planning is then executed to determine the optimal laser scanner location where the best performance on rebar diameter prediction is achieved. For validation, a series of experimental tests including a comparison between the simulated and real scan data were performed on a rebar specimen. The results show that 1) a more than 90% similarity between the simulated and real scan data is achieved with respect to scan density and scan coverage on the rebar surface; and 2) a rebar classification accuracy of 89.3% is achieved, proving the validity of the developed mathematical model and scan planning approach.

COMPLETION OF SPARSE AND PARTIAL POINT CLOUDS OF VEHICLES USING A NOVEL END-TO-END NETWORK

Abstract. Completing the 3D shape of vehicles from real scan data, which aims to estimate the complete geometry of vehicles from partial inputs, acts as a role in the field of remote sensing and autonomous driving. With the recent popularity of deep learning, plenty of data-driven methods have been proposed. However, most of them usually require additional information as prior knowledge for the input, for example, semantic labels and symmetry assumptions. In this paper, we design a novel and end-to-end network, termed as S2U-Net, to achieve the completion of 3D shapes of vehicles from the partial and sparse point clouds. Our network includes two modules of the encoder and the generator. The encoder is designed to extract the global feature of the incomplete and sparse point cloud while the generator is designed to produce fine-grained and dense completion. Specially, we adopt an upsampling strategy to output a more uniform point cloud. Experimental results in the KITTI dataset illustrate our method achieves better performance than the state-of-arts in terms of distribution uniformity and completion quality. Specifically, we improve the translation accuracy by 50.8% and rotation accuracy by 40.6% evaluating completed results with a point cloud registration task.

Topology‐Aware Surface Reconstruction for Point Clouds

AbstractWe present an approach to incorporate topological priors in the reconstruction of a surface from a point scan. We base the reconstruction on basis functions which are optimized to provide a good fit to the point scan while satisfying predefined topological constraints. We optimize the parameters of a model to obtain a likelihood function over the reconstruction domain. The topological constraints are captured by persistence diagrams which are incorporated within the optimization algorithm to promote the correct topology. The result is a novel topology‐aware technique which can (i) weed out topological noise from point scans, and (ii) capture certain nuanced properties of the underlying shape which could otherwise be lost while performing surface reconstruction. We show results reconstructing shapes with multiple potential topologies, compare to other classical surface construction techniques, and show the completion of real scan data.

Vulnerability Selection for Remediation: An Empirical Analysis

Vulnerabilities are security flaws in software and network systems that criminal hackers can exploit to gain an asymmetric advantage. Cyber-Security Operations Centers must routinely triage and patch vulnerabilities in their system(s) to minimize external exposure to attackers. The personnel resources required to address vulnerability remediation tasks are limited and constrained, thus motivating the need for optimization approaches to improve the efficiency of the vulnerability selection process. This paper investigates two different approaches to vulnerability selection for mitigation through (a) Individual Attribute Value Optimization and (b) Multiple Attribute Value Optimization. The former approach presents a methodology that optimizes the selection of vulnerabilities for mitigation with respect to an individual attribute, while the latter approach considers multiple attributes in the vulnerability selection decision-making. Real scan data from a Cyber-Security Operations Center are used to compare the results between the two mathematical approaches. Furthermore, comparisons are made with the results obtained from (a) the actual (baseline) Cyber-Security Operations Center performance, and (b) a vulnerability prioritization algorithm called VULCON that appeared in recent literature.

SMART (SiMulAtion and ReconsTruction) PET: an efficient PET simulation-reconstruction tool

BackgroundPositron-emission tomography (PET) simulators are frequently used for development and performance evaluation of segmentation methods or quantitative uptake metrics. To date, most PET simulation tools are based on Monte Carlo simulations, which are computationally demanding. Other analytical simulation tools lack the implementation of time of flight (TOF) or resolution modelling (RM). In this study, a fast and easy-to-use PET simulation-reconstruction package, SiMulAtion and ReconsTruction (SMART)-PET, is developed and validated, which includes both TOF and RM. SMART-PET, its documentation and instructions to calibrate the tool to a specific PET/CT system are available on Zenodo.SMART-PET allows the fast generation of 3D PET images. As input, it requires one image representing the activity distribution and one representing the corresponding CT image/attenuation map. It allows the user to adjust different parameters, such as reconstruction settings (TOF/RM), noise level or scan duration. Furthermore, a random spatial shift can be included, representing patient repositioning. To evaluate the tool, simulated images were compared with real scan data of the NEMA NU 2 image quality phantom. The scan was acquired as a 60-min list-mode scan and reconstructed with and without TOF and/or RM. For every reconstruction setting, ten statistically equivalent images, representing 30, 60, 120 and 300 s scan duration, were generated. Simulated and real-scan data were compared regarding coefficient of variation in the phantom background and activity recovery coefficients (RCs) of the spheres. Furthermore, standard deviation images of each of the ten statistically equivalent images were compared.ResultsSMART-PET produces images comparable to actual phantom data. The image characteristics of simulated and real PET images varied in similar ways as function of reconstruction protocols and noise levels. The change in image noise with variation of simulated TOF settings followed the theoretically expected behaviour. RC as function of sphere size agreed within 0.3–11% between simulated and actual phantom data.ConclusionsSMART-PET allows for rapid and easy simulation of PET data. The user can change various acquisition and reconstruction settings (including RM and TOF) and noise levels. The images obtained show similar image characteristics as those seen in actual phantom data.

Constructing 3D CSG Models from 3D Raw Point Clouds

AbstractThe Constructive Solid Geometry (CSG) tree, encoding the generative process of an object by a recursive compositional structure of bounded primitives, constitutes an important structural representation of 3D objects. Therefore, automatically recovering such a compositional structure from the raw point cloud of an object represents a high‐level reverse engineering problem, finding applications from structure and functionality analysis to creative redesign. We propose an effective method to construct CSG models and trees directly over raw point clouds. Specifically, a large number of hypothetical bounded primitive candidates are first extracted from raw scans, followed by a carefully designed pruning strategy. We then choose to approximate the target CSG model by the combination of a subset of these candidates with corresponding Boolean operations using a binary optimization technique, from which the corresponding CSG tree can be derived. Our method attempts to consider the minimal description length concept in the point cloud analysis setting, where the objective function is designed to minimize the construction error and complexity simultaneously. We demonstrate the effectiveness and robustness of our method with extensive experiments on real scan data with various complexities and styles.

VULCON

Vulnerability remediation is a critical task in operational software and network security management. In this article, an effective vulnerability management strategy, called VULCON (VULnerability CONtrol), is developed and evaluated. The strategy is based on two fundamental performance metrics: (1) time-to-vulnerability remediation (TVR) and (2) total vulnerability exposure (TVE). VULCON takes as input real vulnerability scan reports, metadata about the discovered vulnerabilities, asset criticality, and personnel resources. VULCON uses a mixed-integer multiobjective optimization algorithm to prioritize vulnerabilities for patching, such that the above performance metrics are optimized subject to the given resource constraints. VULCON has been tested on multiple months of real scan data from a cyber-security operations center (CSOC). Results indicate an overall TVE reduction of 8.97% when VULCON optimizes a realistic security analyst workforce’s effort. Additionally, VULCON demonstrates that it can determine monthly resources required to maintain a target TVE score. As such, VULCON provides valuable operational guidance for improving vulnerability response processes in CSOCs.

New approach for monitoring historic and heritage buildings: Using terrestrial laser scanning and generalised Procrustes analysis

Numerous different techniques and instruments can be used for structural monitoring with different requirements producing different results. For instance, some techniques need to use embedded sensors inside the building, such as geotechnical sensors. However, this method cannot be used for historic and heritage buildings. Other methods can offer high quality, but with a low point density and require fixed stations and targets, such as total stations. In such a case, the location of deformation tends to be known, such as dams, bridges, high-rise buildings, and so forth. Nevertheless, this is not the case for historic and heritage buildings where each block could be subject to deformation. The challenge in such a case is to detect the deformation without any previous knowledge. The aim of this research is to develop a new approach to detect and localise unpredictable deformation. It is based on terrestrial laser scanner measurements and generalised Procrustes analysis techniques to determine deformation vectors, although boxing structure and F-tests are used to detect and localise deformation. In summary, after applying this approach, the whole concerned building is represented as parts, for each of them, the displacement vector and deformation probability are estimated. Validation experiments have shown the capability of the proposed method to detect and localise deformation with magnitude less than noise level in simulated data and of subcentimetre level for ranges up to 10 m in real scan data. Finally, the new approach has been applied to an English Historic site, Bellmanpark Limeklins.

Sub-detector unit timing calibration for brain PET based on L1-norm constraint

A novel sub-detector unit timing calibration method using iterative shrinkage-thresholding (IST) has been developed for a high-resolution brain positron emission tomography (PET) system, which uses a four-layer depth of interaction detector. The timing calibration process is formulated into a linear problem in the proposed method. To obtain a low-variance and robust estimation, an L1-norm constraint is added in the linear equation. Moreover, the detector unit segmentation in the detectors is used to achieve a sub-detector unit level timing calibration. No special source or additional detector element is required for the timing calibration. It is shown that the measured data set from a normal cylindrical phantom filled with radioisotope solution is sufficient for performing a high-precision timing calibration. This method is suitable for PET systems with a large number of crystal elements. In this paper, both simulation and real scanning data were used to demonstrate the effectiveness and robustness of the sub-detector unit level IST method. The timing resolution of a 22Na point source placed at the center of the field of view in the brain PET system was evaluated using different timing calibration methods, along with the optimal selection of detector unit segmentation, to validate the IST method. The timing resolution was improved from 3.34 ns (at full width at half maximum (FWHM)) to 2.54 ns (FWHM) through implementing the IST and the detector unit segmentation.

A Global Parity Measure for Incomplete Point Cloud Data

AbstractShapes with complex geometric and topological features such as tunnels, neighboring sheets, and cavities are susceptible to undersampling and continue to challenge existing reconstruction techniques. In this work we introduce a new measure for point clouds to determine the likely interior and exterior regions of an object. Specifically, we adapt the concept of parity to point clouds with missing data and introduce the parity map, a global measure of parity over the volume. We first examine how parity changes over the volume with respect to missing data and develop a method for extracting topologically correct interior and exterior crusts for estimating a signed distance field and performing surface reconstruction. We evaluate our approach on real scan data representing complex shapes with missing data. Our parity measure is not only able to identify highly confident interior and exterior regions but also localizes regions of missing data. Our reconstruction results are compared to existing methods and we show that our method faithfully captures the topology and geometry of complex shapes in the presence of missing data.

Resolving range ambiguities in high-repetition rate airborne light detection and ranging applications

Correctly determining a measurement range in light detection and ranging instruments, based on time-of-flight measurements on laser pulses, requires the association of each received echo pulse with its causative emitted laser pulse. Without further precautions, this definite association is only possible under specific conditions constraining the usability of range finders and laser scanners with very high measurement rates. We give an introduction to known techniques for avoiding and resolving range ambiguities. The specific disadvantages of these methods led to the development of a technique based on pulse-position modulation (PPM) of the laser pulse train using a pseudorandom noise signal and the subsequent analysis of the impact of the PPM on groups of consecutive range measurements. The error probability of our approach has been evaluated by simulations based on real scan data. The use of a discrete uniform distribution of amplitudes as a modulation signal shows high detection robustness even for difficult terrain like forest areas. The encouraging results of the simulations led to the practical implementation of PPM to our laser scanners and the development of the associated software RiMTA, which resolves range ambiguities without the necessity of a priori information on the target situation and without any user interaction required.

Curvature-Based Environment Description for Robot Navigation Using Laser Range Sensors

This work proposes a new feature detection and description approach for mobile robot navigation using 2D laser range sensors. The whole process consists of two main modules: a sensor data segmentation module and a feature detection and characterization module. The segmentation module is divided in two consecutive stages: First, the segmentation stage divides the laser scan into clusters of consecutive range readings using a distance-based criterion. Then, the second stage estimates the curvature function associated to each cluster and uses it to split it into a set of straight-line and curve segments. The curvature is calculated using a triangle-area representation where, contrary to previous approaches, the triangle side lengths at each range reading are adapted to the local variations of the laser scan, removing noise without missing relevant points. This representation remains unchanged in translation or rotation, and it is also robust against noise. Thus, it is able to provide the same segmentation results although the scene will be perceived from different viewpoints. Therefore, segmentation results are used to characterize the environment using line and curve segments, real and virtual corners and edges. Real scan data collected from different environments by using different platforms are used in the experiments in order to evaluate the proposed environment description algorithm.