Dense 2D Research Articles

Efficient large-scale scene representation with a hybrid of high-resolution grid and plane features

Existing neural radiance fields (NeRF) methods for large-scale scene modeling require days of training using multiple GPUs, hindering their applications in scenarios with limited computing resources. Radiance field models based on explicit dense or hash grid features have been proposed for fast optimization, but their effectiveness has mainly been demonstrated in the context of object-scale scene representation. In this paper, we point out that the low feature resolution in explicit representation is the bottleneck for large-scale unbounded scene representation. To address this problem, we introduce a new and efficient hybrid feature representation for NeRF that fuses the 3D hash-grids and high-resolution 2D dense plane features. Compared with the dense-grid representation, the resolution of a dense 2D plane can be scaled up more efficiently. Based on this hybrid representation, we propose a fast optimization NeRF variant, called GP-NeRF, that achieves better rendering results while maintaining a compact model size. Extensive experiments on multiple large-scale unbounded scene datasets show that our model can converge in 1.5 h using a single GPU while achieving results comparable to or even better than the existing method that requires about one day’s training with 8 GPUs. Our code can be found at https://zyqz97.github.io/GP_NeRF/.

Multi‐Scale Seismic Imaging of the Ridgecrest, CA, Region With Waveform Inversion of Regional and Dense Array Data

AbstractWe develop an inversion procedure for deriving multi‐scale velocity models with waveform inversions of earthquake and ambient noise data at multi‐frequency bands recorded by regional and dense sensor configurations. The method is applied for the area around the 2019 Ridgecrest earthquake rupture zones, utilizing data recorded by regional stations and dense 2D and 1D arrays with station spacings of ∼5 km and ∼100 m, respectively. Starting with regional Vp, Vs models and locations of Ridgecrest aftershocks, the velocity models and event locations are improved iteratively by inversions of waveforms recorded by regional stations and the 2D array, using a minimum spectral element size of ∼600 m. Waveforms from local events recorded by dense 1D arrays across the M7.1 rupture zone with frequencies of up to 10 Hz are used to resolve small‐scale features of the rupture zone and shallow crust with a local spectral element size of 80 m. The refined models provide self‐consistent descriptions of the rupture zone and the shallow crust embedded in the regional structures. The results reveal pronounced low Vs and high Vp/Vs in the M6.4 and M7.1 rupture zones coinciding with concentrations of seismicity, and also around the Garlock fault and in several local basins. We also observe clear velocity contrasts across the Garlock fault with polarity reversals along strike and with depth. The obtained multi‐scale velocity models can be used to improve derivations of earthquake source properties, simulations of dynamic ruptures and ground motions, and the understanding of fault and tectonic processes in the region.

Denoising of Dense Spatial Array Data Using the Curvelet Transform

ABSTRACT To remove background noise from seismic data recorded by spatially dense arrays, we have developed a space-based denoising procedure using the discrete curvelet transform. Based on a detailed statistical characterization of noise coefficients through the empirical cumulative distribution function method within a pre-event time window, signal and noise can be separated effectively by nonlinear thresholding. After synthetic test, we applied this method on data from an industry 3D seismic experiment recorded at an array deployed near Utica, Ohio. The denoising results show good waveform consistency with a significantly enhanced signal-to-noise ratio. Our curvelet approach allows a more computationally efficient spatial–temporal localization analysis of seismic data than conventional curvelet techniques by avoiding the assumption of stationary Gaussian-distributed noise and can be implemented as a complement of time-domain wavelet methods with fewer frequency losses after denoising. This new method provides a fast and convenient way to recover signals from noisy recordings with dense 2D arrays, leading to a considerable improvement in data quality compared with conventional Fourier, wavelet, and curvelet methods. The partitioned seismic signals and noise would yield advanced earth structure imaging, small-event detection, ambient noise tomography, and others.

Landscape evolution during Holocene transgression of a mid‐latitude low‐relief coastal plain: The southern North Sea

AbstractLow‐relief coastal landscapes are at major risk of rising sea levels, as vertical changes in relative sea level have far‐reaching lateral effects. Integration of a dense 2D grid of seismic reflection data with sedimentological and geotechnical data obtained in two offshore wind farm zones allows detailed documentation of postglacial landforms and environmental change over a 1,021 km2 area in the western sector of the southern North Sea. Following the retreat of Last Glacial Maximum ice sheets from the southern North Sea, the resulting postglacial terrestrial landscape provided a surface for peatland formation as climate started to warm and the water table rose in response to relative sea‐level rise. Southward‐draining fluvial networks formed contemporaneously with the peatlands, and remnants of this terrestrial wetland landscape are buried beneath Holocene marine sediments. Distinctive isolated incisional features and discrete widening of fluvial channels that cut through the peats are interpreted as either tidal ponds or relict tidal channels. These features record the evolution of this landscape through the Early Holocene as marine transgression inundated a low‐relief coastal plain. The erosion of the peatlands observed in the cores, the patchy preservation of the organic wetland landscape, and the lack of preserved barrier systems recorded by the seismic reflection data suggest that the rate of relative sea‐level rise outpaced sediment supply during the Late Postglacial and Early Holocene in this area of the southern North Sea. In a regional context, the southward draining river channels contrast to northward fluvial drainage to the North Sea, pointing to a subtle drainage divide in the palaeolandscape and the presence of a low‐relief land bridge separating the North Sea and the English‐Channel/La Manche during the Early Holocene. The documented scenario of rising sea levels combined with decreasing sediment supply in a low‐relief setting is a situation faced by many global deltas and coastlines, which makes the southern North Sea a crucial archive of coastal landscape change.

Angle and Polarization Insensitive Mid-Infrared Optical Filters Using Dense 2D Array of Nanophotonic Cavity Resonators

Angle and Polarization Insensitive Mid-Infrared Optical Filters Using Dense 2D Array of Nanophotonic Cavity Resonators

Separate emission/reception transducers for 3D ultrafast ultrasound imaging

Recent advancements in ultrafast 3D ultrasound imaging have revolutionized the field of echography, enabling its extension to novel applications such as cerebral dynamics, cardiac electrophysiology and the quantitative imaging of intrinsic mechanical properties of tumors. To facilitate these advancements, two primary transducer strategies are employed in 3D imaging. The first involves dense 2D probes equipped with a large number of elements, typically exceeding 1024. The second strategy uses row-column addressing, which simplifies the electronic control of the probe elements. Despite their effectiveness, these methods entail complexities in design and fabrication. Addressing these challenges, our study introduces an innovative transducer configuration that distinctly separates emission and reception functionalities. This separation not only simplifies the overall transducer design but also significantly reduces the system's complexity. A sparse array of PVDF transducers, which have been laser micro-machined to ensure acoustic transparency, is used at the reception. For the emission aspect, we employ a specialized acoustic concentrator to emulate point-like emission. The paper presents the detailed design requirements, assembly process, and the operational principles of this novel transducer. Furthermore, an experimental validation is conducted using a CIRS 040GSE phantom model. This validation is crucial to demonstrate the practical applicability and reliability of our transducer in real-world medical imaging scenarios.

An Open-Access Data Set of Active-Source and Passive-Wavefield DAS and Nodal Seismometer Measurements at the Newberry Florida Site

Abstract This article documents a comprehensive subsurface imaging experiment using seismic waves in a well-studied outdoor laboratory at Newberry, Florida, which is known for significant spatial variability, karstic voids, and underground anomalies. The experiment used approximately two kilometers of distributed acoustic sensing (DAS) fiber-optic cable, forming a dense 2D array of 1920 horizontal-component channels, and a 2D array of 144 SmartSolo three-component nodal seismometers, to sense active-source and passive-wavefield seismic waves. The active-source data were generated using a powerful, triaxial vibroseis shaker truck (T-Rex) and impact sources (accelerated weight drop and an eight-pound sledgehammer) that were simultaneously recorded by both the DAS and nodal seismometers. The vibroseis truck was used to excite the ground in three directions (two horizontal and one vertical) at 260 locations inside and outside the instrumented array, whereas the impact sources were used at 268 locations within the instrumented array. The passive-wavefield data recorded using the nodal seismometers comprised 48 hr of ambient noise collected over a period of four days in four 12-hour time blocks, whereas the passive wavefield data collected using DAS consisted of four hours of ambient noise recordings. This article aims to provide a comprehensive overview of the testing site, experiment layout, the DAS and nodal seismometer acquisition parameters, and implemented raw data processing steps. Although potential use cases, such as surface-wave testing, full-waveform inversion, and ambient noise tomography, are discussed relative to example data, the focus of this article is on documenting this unique data set and presenting its initial data quality rather than on generating subsurface imaging results. The raw and processed data, along with detailed documentation of the experiment and Python tools to aid in visualizing the DAS data set, have been made publicly available.

Reflection seismic imaging of subsurface geologic structures in the Bayan Obo mining area, China

A dense 2D seismic reflection survey was conducted at the Bayan Obo rare-earth-elements-Nb-Fe deposit in China to explore its geologic features and mineral endowment. A composite processing workflow was established for prestack depth imaging of this data set, which involved static correction to compensate for shifts caused by lateral inhomogeneities in near-surface layers and rugged topography. Standard methods were used to attenuate different types of noise and improve the signal-to-noise ratio of the data. A velocity model in the depth domain was obtained from refraction tomography and velocity analysis, and then used in migration. The preliminary interpretation of the final image revealed the mapping of mineralization at its true position in depth, indicating a relatively deep extension (approximately1.5 km). The results demonstrate the potential of seismic imaging as a complementary exploration tool to gravity and electromagnetic methods for mineral exploration purposes in the Bayan Obo mining area, providing a high-resolution seismic image and allowing for depth characterization of the mineral deposits. In addition, the images provide important information for guiding the selection of drilling borehole locations for deep exploration.

Surface energy and stress driven growth of extremely long and high-density ZnO nanowires using a thermal step-oxidation process.

Formation of highly crystalline zinc oxide (ZnO) nanowires with an extremely high aspect ratio (length = 60 μm, diameter = 50 nm) is routinely achieved by introducing an intermediate step-oxidation method during the thermal oxidation process of thin zinc (Zn) films. High-purity Zn was deposited onto clean glass substrates at room temperature using a vacuum-assisted thermal evaporation technique. Afterwards, the as-deposited Zn layers were thermally oxidized under a closed air ambient condition at different temperatures and durations. Structural, morphological, chemical, optical and electrical properties of these oxide layers were investigated using various surface characterization techniques such as X-ray diffraction (XRD), scanning electron microscopy (SEM), Raman spectroscopy, and X-ray photoemission spectroscopy (XPS). It was noticed that the initial thermal oxidation of Zn films usually starts above 400 °C. Homogeneous and lateral growth of the ZnO layer is usually preferred for oxidation at a lower temperature below 500 °C. One-dimensional (1D) asymmetric growth of ZnO started to dominate thermal oxidation above 600 °C. Highly dense 1D ZnO nanowires were specifically observed after prolonged oxidation at 600 °C for 5 hours, followed by short-step oxidation at 700 °C for 30 minutes. However, direct oxidation of Zn films at 700 °C resulted in ZnO nanorod formation. The formation of ZnO nanowires using step-oxidation is explained in terms of surface free energy and compressive stress-driven Zn adatom kinetics through the grain boundaries of laterally grown ZnO seed layers. This simple thermal oxidation process using intermittent step-oxidation was found to be quite unique and very much useful to routinely grow an array of high-density ZnO nanowires. Moreover, these ZnO nanowires showed very high sensitivity and selectivity towards formaldehyde vapour sensing against few other VOCs.

Pseudo-3D cubes from densely spaced subbottom profiles via projection onto convex sets interpolation: An open-source workflow applied to a pockmark field

High-resolution subbottom profiler (SBP) data are commonly recorded during academic or industrial surveys and analyzed as 2D profiles. However, there is growing interest in 3D imaging of the shallow seafloor subsurface, e.g., site surveys prior to the installation of offshore wind farms. Despite ongoing advancements in 3D systems, SBPs continue to be used mainly for 2D profiling. In other seismic imaging applications, “pseudo-3D” seismic cubes have been successfully generated from dense 2D seismic surveys. We developed a novel open-source Python-based workflow to interpolate densely spaced 2D SBP profiles acquired with a hull-mounted parametric SBP into pseudo-3D cubes. This workflow is applied to a study area on the Chatham Rise east of New Zealand’s South Island, comprising numerous seafloor pockmarks and subsurface paleo-pockmarks. Our workflow consists of two stages: (1) processing of individual 2D profiles to compensate for existing vertical offsets and (2) sparse data interpolation by applying the iterative projection onto convex sets algorithm. We correct the residual static effect caused by the sea state during acquisition, compensate for the varying tidal elevations, and apply vertical time shifts to individual profiles to correct misties at profile intersections. The interpolation of the binned sparse 3D cube is performed in the frequency domain using the fast Fourier transform and executed in parallel processes, which significantly increases the computational efficiency. We validate our workflow by testing different bin sizes, sparse transforms, thresholding approaches, and varying total iterations, which noticeably impact the resulting interpolation quality. Our developed workflow generates pseudo-3D subsurface images from densely spaced SBP profiles, with numerous potential applications for academic and industrial surveys.

FusionNet: An End-to-End Hybrid Model for 6D Object Pose Estimation

In this study, we propose a hybrid model for Perspective-n-Point (PnP)-based 6D object pose estimation called FusionNet that takes advantage of convolutional neural networks (CNN) and Transformers. CNN is an effective and potential tool for feature extraction, which is considered the most popular architecture. However, CNN has difficulty in capturing long-range dependencies between features, and most CNN-based models for 6D object pose estimation are bulky and heavy. To address these problems, we propose a lighter-weight CNN building block with attention, design a Transformer-based global dependency encoder, and integrate them into a single model. Our model is able to extract dense 2D–3D point correspondences more accurately while significantly reducing the number of model parameters. Followed with a PnP header that replaces the PnP algorithm for general end-to-end pose estimation, our model showed better or highly competitive performance in pose estimation compared with other state-of-the-art models in experiments on the LINEMOD dataset.

Triangle-Mesh-Rasterization-Projection (TMRP): An Algorithm to Project a Point Cloud onto a Consistent, Dense and Accurate 2D Raster Image.

The projection of a point cloud onto a 2D camera image is relevant in the case of various image analysis and enhancement tasks, e.g., (i) in multimodal image processing for data fusion, (ii) in robotic applications and in scene analysis, and (iii) for deep neural networks to generate real datasets with ground truth. The challenges of the current single-shot projection methods, such as simple state-of-the-art projection, conventional, polygon, and deep learning-based upsampling methods or closed source SDK functions of low-cost depth cameras, have been identified. We developed a new way to project point clouds onto a dense, accurate 2D raster image, called Triangle-Mesh-Rasterization-Projection (TMRP). The only gaps that the 2D image still contains with our method are valid gaps that result from the physical limits of the capturing cameras. Dense accuracy is achieved by simultaneously using the 2D neighborhood information (rx,ry) of the 3D coordinates in addition to the points P(X,Y,V). In this way, a fast triangulation interpolation can be performed. The interpolation weights are determined using sub-triangles. Compared to single-shot methods, our algorithm is able to solve the following challenges. This means that: (1) no false gaps or false neighborhoods are generated, (2) the density is XYZ independent, and (3) ambiguities are eliminated. Our TMRP method is also open source, freely available on GitHub, and can be applied to almost any sensor or modality. We also demonstrate the usefulness of our method with four use cases by using the KITTI-2012 dataset or sensors with different modalities. Our goal is to improve recognition tasks and processing optimization in the perception of transparent objects for robotic manufacturing processes.

Magnetic saturation enhancement of gold-capped nickel nanorods

Nickel nanorods (NRs) capped with gold (Au/Ni) were grown into porous anodic aluminum oxide templates and subsequently transferred onto Au/Si (100) substrates. A high dense 2D array of Ni and Au/Ni nanorods was analyzed by vibrating sample magnetometry; it was found that an increase in 14.8% of the magnetic moment following the deposition of Au caps. In order to further investigate this phenomenon, the magnetic distribution of Au/Ni nanorods was studied by off-axis electron holography. The magnetization and induction strengths were evaluated to be 4.7 × 105 A/m and 0.62 T, respectively, which is equivalent to magnetometry measurements of the Ni NR arrays. Remarkably, a vortex state configuration was imaged in the Au segment by the retrieved magnetic phase of the electron holograms under free lens conditions of the transmission electron microscope column. It was concluded that the magnetic distribution in the Au segment is associated with a ferromagnetic coupling with Ni and correlated with the magnetometry measurements.

LWSIS: LiDAR-Guided Weakly Supervised Instance Segmentation for Autonomous Driving

Image instance segmentation is a fundamental research topic in autonomous driving, which is crucial for scene understanding and road safety. Advanced learning-based approaches often rely on the costly 2D mask annotations for training. In this paper, we present a more artful framework, LiDAR-guided Weakly Supervised Instance Segmentation (LWSIS), which leverages the off-the-shelf 3D data, i.e., Point Cloud, together with the 3D boxes, as natural weak supervisions for training the 2D image instance segmentation models. Our LWSIS not only exploits the complementary information in multimodal data during training but also significantly reduces the annotation cost of the dense 2D masks. In detail, LWSIS consists of two crucial modules, Point Label Assignment (PLA) and Graph-based Consistency Regularization (GCR). The former module aims to automatically assign the 3D point cloud as 2D point-wise labels, while the atter further refines the predictions by enforcing geometry and appearance consistency of the multimodal data. Moreover, we conduct a secondary instance segmentation annotation on the nuScenes, named nuInsSeg, to encourage further research on multimodal perception tasks. Extensive experiments on the nuInsSeg, as well as the large-scale Waymo, show that LWSIS can substantially improve existing weakly supervised segmentation models by only involving 3D data during training. Additionally, LWSIS can also be incorporated into 3D object detectors like PointPainting to boost the 3D detection performance for free. The code and dataset are available at https://github.com/Serenos/LWSIS.

Formation of Porosity toward Acetylene upon Vitrification of Non-porous Photochromic Coordination Polymer Crystals

Designing permanent microporosity in a glassy state of coordination polymer (CP) and metal–organic framework (MOF) has been challenging due to densification of the structures upon vitrification. Here, we propose a new approach to create porosity in CP/MOF glasses derived from nonporous 1D crystal structures. Glassy states of two dense 1D crystal structures [Ag2(cis-dbe) (CF3SO3)2] and [Ag2(cis-dbe) (CF3CO2)2], where cis-dbe = cis-1,2-dicyano-1,2-bis(2,4,5-trimethyl-3-thienyl)ethene, were prepared by ball milling. Glass structures and dynamics were characterized by synchrotron X-ray, solid-state NMR, and thermal analyses. Loosened assembly of 1D chains through glass formation enabled representation of permanent microporosity and resulted in selective gas sorption behaviors of acetylene over CO2. It elucidated that the property of gas uptake is affected by the glass dynamics and interaction with anions. Reversible photochromism of cis-dbe is preserved in the glasses, leading to multi-functional glass materials.

Distributed swarm collision avoidance based on angular calculations

Collision avoidance is one of the most important topics in the robotics field. In this problem, the goal is to move the robots from initial locations to target locations such that they follow the shortest non-colliding paths in the shortest time and with the least amount of energy. Robot navigation among pedestrians is an example application of this problem which is the focus of this paper. This paper presents a distributed and real-time algorithm for solving collision avoidance problems in dense and complex 2D and 3D environments. This algorithm uses angular calculations to select the optimal direction for the movement of each robot and it has been shown that these separate calculations lead to a form of cooperative behavior among agents. We evaluated the proposed approach on various simulation and experimental scenarios and compared the results with ORCA one of the most important algorithms in this field. The results show that the proposed approach is at least 25% faster than ORCA while is also more reliable. The proposed method is shown to enable fully autonomous navigation of a swarm of Crazyflies.

Impact of illumination on quantum lifetime in selectively doped GaAs single quantum wells with short-period AlAs/GaAs superlattice barriers

Impact of illumination on high-mobility dense 2D electron gas in selectively doped single GaAs quantum well with short-period AlAs/GaAs superlattice barriers at T=4.2 K in magnetic fields B<2 T has been studied. It was demonstrated that illumination at low temperatures gives rise to enhancement of electron density, mobility and quantum lifetime in studied heterostructures. The enhancement of quantum lifetime after illumination for single GaAs quantum well with modulated superlattice doping had been explained as consequence of decrease in effective concentration of remote ionized donors. Keywords: persistent photoconductivity, quantum lifetime, anisotropic mobility, superlattice barriers.

VPFNet: Improving 3D Object Detection With Virtual Point Based LiDAR and Stereo Data Fusion

It has been well recognized that fusing the complementary information from depth-aware LiDAR point clouds and semantic-rich stereo images would benefit 3D object detection. Nevertheless, it is non-trivial to explore the inherently unnatural interaction between sparse 3D points and dense 2D pixels. To ease this difficulty, the recent approaches generally project the 3D points onto the 2D image plane to sample the image data and then aggregate the data at the points. However, these approaches often suffer from the mismatch between the resolution of point clouds and RGB images, leading to sub-optimal performance. Specifically, taking the sparse points as the multi-modal data aggregation locations causes severe information loss for high-resolution images, which in turn undermines the effectiveness of multi-sensor fusion. In this paper, we present <monospace>VPFNet</monospace>&#x2014;a new architecture that cleverly aligns and aggregates the point cloud and image data at the &#x201C;virtual&#x201D; points. Particularly, with their density lying between that of the 3D points and 2D pixels, the virtual points can nicely bridge the resolution gap between the two sensors, and thus preserve more information for processing. Moreover, we also investigate the data augmentation techniques that can be applied to both point clouds and RGB images, as the data augmentation has made non-negligible contribution towards 3D object detectors to date. We have conducted extensive experiments on KITTI dataset, and have observed good performance compared to the state-of-the-art methods. Remarkably, our <monospace>VPFNet</monospace> achieves 83.21&#x0025; moderate <inline-formula><tex-math notation="LaTeX">$AP_{3D}$</tex-math></inline-formula> and 91.86&#x0025; moderate <inline-formula><tex-math notation="LaTeX">$AP_{BEV}$</tex-math></inline-formula> on the KITTI test set. The network design also takes computation efficiency into consideration &#x2013; we can achieve a FPS of 15 on a single NVIDIA RTX 2080Ti GPU. The source code is available at <uri>https://github.com/zhukevkesky/VPFNet.git</uri>https://github.com/zhukevkesky/VPFNet.git.

Влияние подсветки на квантовое время жизни в селективно-легированных одиночных GaAs квантовых ямах с короткопериодными AlAs/GaAs-сверхрешеточными барьерами

Impact of illumination on high-mobility dense 2D electron gas in selectively doped single GaAs quantum well with short-period AlAs/GaAs superlattice barriers at T = 4.2 K in magnetic fields B &lt; 2 T has been studied. It was demonstrated that illumination at low temperatures gives rise to enhancement of electron density, mobility and quantum lifetime in studied heterostructures. The enhancement of quantum lifetime after illumination for single GaAs quantum well with modulated superlattice doping had been explained as consequence of decrease in effective concentration of remote ionized donors.

A Free-Standing Polyaniline/Silicon Nanowire Forest as the Anode for Lithium-ion Batteries.

Despite its high theoretical capacity, silicon anode has limited intrinsic conductivity and experiences significant volume changes during charge-discharge. To overcome these issues, facile metal-assisted chemical etching and in-situ polymerization of aniline are employed to produce a dense 1D polyaniline/silicon nanowire forest without noticeable agglomeration as a free-standing anode for lithium-ion batteries. This hybrid electrode possesses high cycling performance, delivering a stable capacity capped at 2 mAh cm-2 for 346 cycles of charge-discharge. Maximum capacity of 2 mAh cm-2 is also achievable at high-rate cell testing of 2 mA cm-2 , which cannot be obtained by the anode with plain silicon wafer and silicon nanowire only. The introduction of polyaniline on the silicon nanowire is shown to reduce the solid electrolyte interface (SEI) resistance, stabilize the SEI layer, further alleviate the effect of volume changes, and boost the conductivity of the hybrid anode, resulting in the high electrochemical performance of the anode.