Over the past decade deep learning has driven progress in 2D image understanding. Despite these advancements, techniques for automatic 3D sensed data understanding, such as point clouds, is ...comparatively immature. However, with a range of important applications from indoor robotics navigation to national scale remote sensing there is a high demand for algorithms that can learn to automatically understand and classify 3D sensed data. In this paper we review the current state-of-the-art deep learning architectures for processing unstructured Euclidean data. We begin by addressing the background concepts and traditional methodologies. We review the current main approaches, including RGB-D, multi-view, volumetric and fully end-to-end architecture designs. Datasets for each category are documented and explained. Finally, we give a detailed discussion about the future of deep learning for 3D sensed data, using literature to justify the areas where future research would be most valuable.
Robust and reliable automatic building detection and segmentation from aerial images/point clouds has been a prominent field of research in remote sensing, computer vision and point cloud processing ...for a number of decades. One of the largest issues associated with deep learning methods is the high quantity of data required for training. To help address this we present a method to improve public GIS building footprint labels by using Morphological Geodesic Active Contours (MorphGACs). We demonstrate by improving the quality of building footprint labels for detection and semantic segmentation, more robust and reliable models can be obtained. We evaluate these methods over a large UK-based dataset of 24556 images containing 169835 building instances. This is achieved by training several Mask/Faster R-CNN and RetinaNet deep convolutional neural networks. Networks are supplied with both RGB and fused RGB-lidar data. We offer quantitative analysis on the benefits of the inclusion of depth data for building segmentation. By employing both methods we achieve a detection accuracy of 0.92 (mAP@0.5) and segmentation f1 scores of 0.94 over a 4911 test images ranging from urban to rural scenes.
The Sonic Kayak is a musical instrument used to investigate nature and developed during open hacklab events. The kayaks are rigged with underwater environmental sensors, which allow paddlers to hear ...real-time water temperature sonifications and underwater sounds, generating live music from the marine world. Sensor data is also logged every second with location, time and date, which allows for fine-scale mapping of water temperatures and underwater noise that was previously unattainable using standard research equipment. The system can be used as a citizen science data collection device, research equipment for professional scientists, or a sound art installation in its own right.
A catalogue of hidden momenta Griffiths, David J.
Philosophical transactions - Royal Society. Mathematical, Physical and engineering sciences/Philosophical transactions - Royal Society. Mathematical, physical and engineering sciences,
10/2018, Letnik:
376, Številka:
2134
Journal Article
Recenzirano
Odprti dostop
Electromagnetic fields carry momentum: . But if the centre of energy of a (localized) system is at rest, its total momentum must be zero. The compensating term has come to be called 'hidden' ...momentum: Ph = − Pem. It is (typically) ordinary mechanical momentum, relativistic in nature, and is 'hidden' only in the sense that it is not associated with motion of the system as a whole-only with that of its constituent parts. This article develops a catalogue of field momenta and hidden momenta for ideal electric and magnetic dipoles-both the 'standard' variety made from electric charges and currents and the 'anomalous' variety made from hypothetical magnetic monopoles and their currents-in the presence of electric and magnetic fields (which themselves may be produced by 'standard' or 'anomalous' sources).
This article is part of the theme issue 'Celebrating 125 years of Oliver Heaviside's 'Electromagnetic Theory''.
While significant progress has been made in the last few years, the reliability of Additively Manufactured (AM) parts is often less than desirable as they suffer from manufacturing defects and hence ...subpar strength and fatigue life. To address this challenge, numerical methods are sought to provide insight into the process and help accelerate progress in raising the quality of AM parts. In metal AM applications, assessing the amount of unfused powder, melt pool volumes, and metallurgical phase transformations is often of interest. In this work, we introduce a generic framework for assessing metallurgical phase transformations, building on a previously-developed general simulation framework for predicting temperature evolution, distortions, and residual stresses. Experimental work was conducted to validate numerical predictions and included temperature measurements, EBSD/XRD microstructural examinations. Additional test data from the published literature was used to validate melt pool sizes and unfused powder predictions. We conclude that the continuum-level internal state variable approach proposed here can be calibrated with a reasonable amount of effort and can be used directly to link processing conditions to microstructural features. Upcoming work investigates the influence of microstructure features on mechanical performance to address the process-structure-property-performance relationship in SLM additive manufacturing.
Display omitted
•A generic framework for simulation of Additive Manufactured parts•Linking thermal model to microstructural features•Model validation in Ti-6Al-4 V and Steel 5140•Experiments and simulations of Selective Laser Melting processed Ti-6Al-4 V samples•Microstructure assessment of as-built and heat-treated Ti-6Al-4 V samples
AccessLabs are workshops with two simultaneous motivations, achieved through direct citizen-scientist pairings: (1) to decentralise research skills so that a broader range of people are able to ...access/use scientific research, and (2) to expose science researchers to the difficulties of using their research as an outsider, creating new open access advocates. Five trial AccessLabs have taken place for policy makers, media/journalists, marine sector participants, community groups, and artists. The act of pairing science academics with local community members helps build understanding and trust between groups at a time when this relationship appears to be under increasing threat from different political and economic currents in society. Here, we outline the workshop motivations, format, and evaluation, with the aim that others can build on the methods developed.
Studying ecological and evolutionary processes in the natural world often requires research projects to follow multiple individuals in the wild over many years. These projects have provided ...significant advances but may also be hampered by needing to accurately and efficiently collect and store multiple streams of the data from multiple individuals concurrently. The increase in the availability and sophistication of portable computers (smartphones and tablets) and the applications that run on them has the potential to address many of these data collection and storage issues. In this paper we describe the challenges faced by one such long-term, individual-based research project: the Banded Mongoose Research Project in Uganda. We describe a system we have developed called Mongoose 2000 that utilises the potential of apps and portable computers to meet these challenges. We discuss the benefits and limitations of employing such a system in a long-term research project. The app and source code for the Mongoose 2000 system are freely available and we detail how it might be used to aid data collection and storage in other long-term individual-based projects.
Bridges are a critical piece of infrastructure in the network of road and rail transport system. Many of the bridges in Norway (in Europe) are at the end of their lifespan, therefore regular ...inspection and maintenance are critical to ensure the safety of their operations. However, the traditional inspection procedures and resources required are so time consuming and costly that there exists a significant maintenance backlog. The central thrust of this paper is to demonstrate the significant benefits of adapting a Unmanned Aerial Vehicle (UAV)-assisted inspection to reduce the time and costs of bridge inspection and established the research needs associated with the processing of the (big) data produced by such autonomous technologies. In this regard, a methodology is proposed for analysing the bridge damage that comprises three key stages, (i) data collection and model training, where one performs experiments and trials to perfect drone flights for inspection using case study bridges to inform and provide necessary (big) data for the second key stage, (ii) 3D construction, where one built 3D models that offer a permanent record of element geometry for each bridge asset, which could be used for navigation and control purposes, (iii) damage identification and analysis, where deep learning-based data analytics and modelling are applied for processing and analysing UAV image data and to perform bridge damage performance assessment. The proposed methodology is exemplified via UAV-assisted inspection of Skodsberg bridge, a 140 m prestressed concrete bridge, in the Viken county in eastern Norway.