Road networks are essential elements of a community's infrastructure and need regular inspection. Present practice requires traffic interruptions and safety risks for inspectors. The road detection system based on vehicle-mounted lasers is also quite mature, offering advantages such as high-precision defect detection, high automation, and fast detection speed. However, it does have drawbacks such as high equipment procurement and maintenance costs, limited flexibility, and insufficient coverage range. Therefore, this paper proposes a low-cost unmanned aerial vehicle (UAV)-based alternative using imagery for automatic road pavement inspection focusing on pothole detection and classification. A slicing-based method, entitled the Pavement Pothole Detection Algorithm, is applied to the imagery after it is converted into a three-dimensional point cloud. When compared with manually extracted results, the proposed UAV-structure-from-motion (SfM) method and the associated algorithm achieved 0.01 m level accuracy for pothole depth detection and maximum errors of 0.0053 m3 in volume evaluation for cases studies of both a road and a bridge deck.

@en

Imagery from Unmanned Aerial Vehicles can be used to generate three-dimensional (3D) point cloud models. However, final data quality is impacted by the flight altitude, camera angle, overlap rate, and data processing strategies. Typically, both overview images and redundant close-range images are collected, which significantly increases the data collection and processing time. To investigate the relationship between input resources and output quality, a suite of seven metrics is proposed including total points, average point density, uniformity, yield rate, coverage, geometry accuracy, and time efficiency. When applied in the field to a full-scale structure, the UAV altitude and camera angle most strongly affected data density and uniformity. A 66% overlapping was needed for successful 3D reconstruction. Conducting multiple flight paths improved local geometric accuracy better than increasing the overlapping rate. The highest coverage was achieved at 77% due to the formation of semi-irregular gridded gaps between point groups as an artefact of the Structure from Motion process. No single set of flight parameters was optimal for every data collection goal. Hence, understanding flight path parameter impacts is crucial to optimal UAV data collection.

@en

Most deep learning (DL) methods that are not end-to-end use several multi-scale and multi-type hand-crafted features that make the network challenging, more computationally intensive and vulnerable to overfitting. Furthermore, reliance on empirically-based feature dimensionality reduction may lead to misclassification. In contrast, efficient feature management can reduce storage and computational complexities, builds better classifiers, and improves overall performance. Principal Component Analysis (PCA) is a well-known dimension reduction technique that has been used for feature extraction. This paper presents a two-step PCA based feature extraction algorithm that employs a variant of feature-based PointNet (Qi et al., 2017a) for point cloud classification. This paper extends the PointNet framework for use on large-scale aerial LiDAR data, and contributes by (i) developing a new feature extraction algorithm, (ii) exploring the impact of dimensionality reduction in feature extraction, and (iii) introducing a non-end-to-end PointNet variant for per point classification in point clouds. This is demonstrated on aerial laser scanning (ALS) point clouds. The algorithm successfully reduces the dimension of the feature space without sacrificing performance, as benchmarked against the original PointNet algorithm. When tested on the well-known Vaihingen data set, the proposed algorithm achieves an Overall Accuracy (OA) of 74.64% by using 9 input vectors and 14 shape features, whereas with the same 9 input vectors and only 5PCs (principal components built by the 14 shape features) it actually achieves a higher OA of 75.36% which demonstrates the effect of efficient dimensionality reduction.

@en

This paper presents the work carried out on a collaborative tripartite project between the USA, Republic of Ireland and Northern Ireland to create and investigate the design, development and testing of a new class of intermeshed steel connections (ISCs) that do not rely on field welding and minimise bolting, thus targeting the facilitation of fast disassembly of steel structures and material reuse. This research took advantage of fully automated, precise, advanced manufacturing cutting technologies (e.g. laser, waterjet and high-definition plasma cutting) to achieve a connection method in steel that previously was only possible in materials such as timber, with the potential to revolutionise the steel construction industry. The paper outlines the ongoing research work by the collaborative team, focusing on the design, fabrication, finite-element analysis (FEA) and scaled experimental testing of side ISCs for the flanges of open sections, which included the use of state-of-the-art digital image correlation technology for non-contact measurements. A simplified connection design procedure is presented based on yielding of the side plates. This design procedure is refined based on the results of experimental testing and FEA of the local axial behaviour of the flange connection, addressing stress concentrations in the flange, fabrication tolerances and material overstrength.

@en

This paper describes an investigation into the use of advanced manufacturing techniques for the creation of a new class of intermeshed steel connections that rely on neither welding nor bolting. The project detailed herein lays the groundwork to transform the steel building construction industry by advancing the underlying science and engineering precepts for intermeshed connections created from precise, volumetric cutting. The proposed system enhances the integration between design, fabrication, and installation. Fully automated, precise, volumetric cutting of open steel sections poses challenges regarding the load-Transfer mechanisms and failure modes for intermeshed connections. Implementation of the intermeshed connection would cause a discontinuity in the beam; therefore, the effects of such a configuration on the behavior of the steel frame are investigated in the current paper. Load resistance and design of these connections are also explored with physical tests and finite element modeling to investigate the mechanics of intermeshed connections, including stress and strain concentrations, fracture and failure modes, and connection geometry optimization.

@en

When aerial laser scanning (ALS) is deployed with targeted flight path planning, urban scenes can be captured in points clouds with both high vertical and horizontal densities to support a new generation of urban analysis and applications. As an example, this paper proposes a hierarchical method to automatically extract data points describing road edges, which are then used for reconstructing road edges and identifying accessible passage areas. The proposed approach is a cell-based method consisting of 3 main steps: (1) filtering rough ground points, (2) extracting cells containing data points of the road curb, and (3) eliminating incorrect road curb segments. The method was tested on a pair of 100 m × 100 m tiles of ALS data of Dublin Ireland's city center with a horizontal point density of about 325 points/m2. Results showed the data points of the road edges to be extracted properly for locations appearing as the road edges with the average distance errors of 0.07 m and the ratio between the extracted road edges and the ground truth by 73.2%.

@en

Current processes that create Building Energy Performance Simulation (BEPS) models are time consuming and costly, primarily due to the extensive manual inputs required for model population. In particular, generation of geometric inputs for existing building models requires significant manual intervention due to the absence, or outdated nature of available data or digital measurements. Additionally, solutions based on Building Information Modelling (BIM) also require high quality and precise geometrically-based models, which are not typically available for existing buildings. As such, this work introduces a semi-automated BEPS input solution for existing building exteriors that can be integrated with other related technologies (such as BIM or CityGML) and deployed across an entire building stock. Within the overarching approach, a novel sub-process automatically transforms a point cloud obtained from a terrestrial laser scanner into a representation of a building's exterior façade geometry as input data for a BEPS engine. Semantic enrichment is performed manually. This novel solution extends two existing approaches: (1) an angle criterion in boundary detection and (2) a voxelisation representation to improve performance. The use of laser scanning data reduces temporal costs and improves input accuracy for BEPS model generation of existing buildings. The approach is tested herein on two example cases. Vertical and horizontal accuracies of 1% and 7% were generated, respectively, when compared against independently produced, measured drawings. The approach showed variation in accuracy of model generation, particularly for upper floors of the test case buildings. However, the energy impacts resulting from these variations represented less than 1% of the energy consumption for both cases.@en