Entire Clouds Research Articles

A robust assessment method of point cloud quality for enhancing 3D robotic scanning

Point clouds are widely used to construct models of workpieces using 3D scanners, especially where high-quality robotic and automatic 3D scanning is required in industries and manufacturing. In recent years, Point Cloud Quality Assessment (PCQA) has garnered increasing attention as it provides quality scores for entire point clouds, addressing issues such as downsampling and compression distortions. However, current PCQA methods cannot provide specific and local quality scores, which are necessary to facilitate rescanning and recompletion in 3D robotic scanning. Additionally, 3D data produced by robotic view-planning algorithms are usually considered the final result, where PCQA is typically not involved. In this paper, we bridge the gap between PCQA methods and practical robotic 3D scanning. We propose a no-reference PCQA method that recognizes sparse regions during 3D scanning, providing both local and overall quality scores. Unlike traditional methods that primarily consider density as a key metric, our method assumes that an expected 3D scan will have a uniformly distributed point cloud on surfaces. We analyze the quality of points by using geometric information from surfaces fitted to these points, which are mapped to a 2D distribution based on specified distances and angles. We conducted experiments on various datasets, including both synthetic and public datasets, to evaluate the accuracy and robustness of our method. The results show that our method can represent the quality on surfaces more accurately and robustly than density calculation methods. Additionally, it outperforms most existing PCQA methods in scenarios of downsampling, which is a common challenge in high-quality 3D scanning applications. The performance of our quality enhancement experiments on practical 3D scanning, conducted towards the end of our study, demonstrates significant potential for real-world applications. The related code is released at https://github.com/leihui6/PQA.

Adaptive LiDAR Reconstruction by Convolutional Compressive Sensing Network and Multivariate Empirical Mode Decomposition

With the recent advancement of LiDAR data processing pipelines integrated with signal processing techniques, it is now possible to rapidly acquire, process, and infer multidimensional information of 3D surroundings. This work proposes a novel compressive sensing (CS) framework (named MEMD-CSNet), based on Multivariate Empirical Mode Decomposition (MEMD) for adaptive 3D reconstruction of airborne urban LiDAR point clouds. EMD is a data-driven transform for decomposing the input data into Intrinsic Mode Functions (IMFs), capturing the implicit features in terms of variation in spatial frequency. Our approach generates a new filtered 3D signal constituting an optimal subset of IMFs based on the understanding of their statistical significance. Further, this filtered signal hierarchically extracts the overall shape in the last IMF (Residual function). Our proposed MEMD-CSNet is a deep convolutional network that (1) computes a multivariate EMD on input raw 3D LiDAR point clouds for generating data-dependent feature-rich filtered signal and (2) additionally augments this signal with convolutional features, thus learning a data-centric representation function (Ψ) for transforming the input into a low-dimensional sparse basis. We present extensive qualitative and quantitative analysis using the proposed MEMD-CSNet for a set of chosen CS measurement ratios. MEMD-CSNet achieves a maximum reconstruction PSNR of 108.22 dB (approx. 30 dB/1.36x better than without using MEMD) for reconstruction fidelity on selected urban LiDAR scenes. We envisage that our proposed MEMD-CSNet has a high potential (as a data-centric framework) for generating data-driven sparse feature maps and reconstructing entire 3D LiDAR point clouds from a few sparse measurements.

Improved Registration Algorithm Based on Double Threshold Feature Extraction and Distance Disparity Matrix.

Entire surface point clouds in complex objects cannot be captured in a single direction by using noncontact measurement methods, such as machine vision; therefore, different direction point clouds should be obtained and registered. However, high efficiency and precision are crucial for registration methods when dealing with huge number of point clouds. To solve this problem, an improved registration algorithm based on double threshold feature extraction and distance disparity matrix (DDM) is proposed in this study. Firstly, feature points are extracted with double thresholds using normal vectors and curvature to reduce the number of points. Secondly, a fast point feature histogram is established to describe the feature points and obtain the initial corresponding point pairs. Thirdly, obviously wrong corresponding point pairs are eliminated as much as possible by analysing the Euclidean invariant features of rigid body transformation combined with the DDM algorithm. Finally, the sample consensus initial alignment and the iterative closest point algorithms are used to complete the registration. Experimental results show that the proposed algorithm can quickly process large data point clouds and achieve efficient and precise matching of target objects. It can be used to improve the efficiency and precision of registration in distributed or mobile 3D measurement systems.

Fast and Accurate Registration of the Proximal Femurs in Bilateral Hip Joint Images by Using the Random Sub-Sample Points

Background and ObjectiveAs is known, point clouds representing the objects are frequently used in object registration. Although the objects can be registered by using all the points in the corresponding point clouds of the objects, the registration process can also be achieved with a smaller number of the landmark points selected from the entire point clouds of the objects. This paper introduces a research study focusing on the fast and accurate rigid registration of the bilateral proximal femurs in bilateral hip joint images by using the random sub-sample points. For this purpose, Random Point Sub-sampling (RPS) was analyzed and the reduced point sets were used for an accurate registration of the bilateral proximal femurs in coronal hip joint magnetic resonance imaging (MRI) slices. MethodsIn registration, bilateral proximal femurs in MRI slices were registered rigidly by performing a process consisting of three main phases named as MR image preprocessing, proximal femur registration over the random sub-sample points and MR image postprocessing. In the stage of the MR image preprocessing, segmentation maps of the bilateral proximal femurs are obtained as region of interest (RoI) images from the entire MRI slices and then, the edge maps of the segmented proximal femurs are extracted. In the registration phase, the edge maps describing the proximal femur surfaces are represented as point clouds initially. Thereafter, the RPS is performed on the proximal femur point clouds and the number of points representing the proximal femurs is reduced at different ratios. For the registration of the point clouds, the Iterative Closest Point (ICP) algorithm is performed on the reduced sets of points. Finally, the registration procedures are completed by performing MR image postprocessing on the registered proximal femur images. ResultsIn performance evaluation tests performed on healthy and pathological proximal femurs in 13 bilateral coronal hip joint MRI slices of 13 Legg-Calve-Perthes disease (LCPD) patients, bilateral proximal femurs were successfully registered with very small error rates by using the reduced set of points obtained via the RPS and promising results were achieved. The minimum error rate was observed at RPS rate of 30% as the value of 0.41 (±0.31)% on all over the bilateral proximal femurs evaluated. When the range of RPS rate of 20-30% is considered as the reference, the elapsed time in registration can be reduced by almost 30-40% compared to the case where all the proximal femur points were included in registration. Additionally, it was observed that the RPS rate should be selected as at least 25% to achieve a successful registration with an error rate below 1%. ConclusionIt was concluded from the observed results that a more successful and faster registration can be accomplished by selecting fewer points randomly from the point sets of proximal femurs instead of using all the points describing the proximal femurs. Not only an accurate registration with low error rates was performed, but also a faster registration process was performed by means of the limited number of points that are sub-sampled randomly from the whole point sets.

Effects of initial density profiles on massive star cluster formation in giant molecular clouds

ABSTRACT We perform a suite of hydrodynamic simulations to investigate how initial density profiles of giant molecular clouds (GMCs) affect their subsequent evolution. We find that the star formation duration and integrated star formation efficiency of the whole clouds are not sensitive to the choice of different profiles but are mainly controlled by the interplay between gravitational collapse and stellar feedback. Despite this similarity, GMCs with different profiles show dramatically different modes of star formation. For shallower profiles, GMCs first fragment into many self-gravitation cores and form subclusters that distributed throughout the entire clouds. These subclusters are later assembled ‘hierarchically’ to central clusters. In contrast, for steeper profiles, a massive cluster is quickly formed at the centre of the cloud and then gradually grows its mass via gas accretion. Consequently, central clusters that emerged from clouds with shallower profiles are less massive and show less rotation than those with the steeper profiles. This is because (1) a significant fraction of mass and angular momentum in shallower profiles is stored in the orbital motion of the subclusters that are not able to merge into the central clusters, and (2) frequent hierarchical mergers in the shallower profiles lead to further losses of mass and angular momentum via violent relaxation and tidal disruption. Encouragingly, the degree of cluster rotations in steeper profiles is consistent with recent observations of young and intermediate-age clusters. We speculate that rotating globular clusters are likely formed via an ‘accretion’ mode from centrally concentrated clouds in the early Universe.

Evidence of Cosmic-Ray Excess from Local Giant Molecular Clouds

Abstract We report the analysis of the Fermi-Large Area Telescope data from six nearby giant molecular clouds (MCs) belonging to the Gould Belt and the Aquila Rift regions. The high statistical γ-ray spectra above 3 GeV well described by power laws make it possible to derive precise estimates of the cosmic-ray (CR) distribution in the MCs. The comparison of γ-ray spectra of Taurus, Orion A, and Orion B clouds with the model expected from Alpha Magnetic Spectrometer (AMS-02) CR measurements confirms these clouds as passive clouds, immersed in an AMS-02-like CR spectrum. A similar comparison of Aquila Rift, Rho Oph, and Cepheus spectra yields significant deviation in both spectral indices and absolute fluxes, which can imply an additional acceleration of CRs throughout the entire clouds. Besides, the theoretical modeling of the excess γ-ray spectrum of these clouds, assuming π 0-decay interaction of CRs in the cloud, gives a considerable amount of an enhanced CR energy density and it shows a significant deviation in spectral shapes compared to the average AMS-02 CR spectrum between 30 GeV and 10 TeV. We suggest that this variation in the CR spectrum of Cepheus could be accounted for by an efficient acceleration in the shocks of winds of OB associations, while in Rho Oph, similar acceleration can be provided by multiple T-Tauri stars populated in the whole cloud. In the case of Aquila Rift, the excess in absolute CR flux could be related to an additional acceleration of CRs by supernova remnants or propagation effects in the cloud.

HOLISTIC PARAMETRIC RECONSTRUCTION OF BUILDING MODELS FROM POINT CLOUDS

Abstract. Building models are conventionally reconstructed by building roof points via planar segmentation and then using a topology graph to group the planes together. Roof edges and vertices are then mathematically represented by intersecting segmented planes. Technically, such solution is based on sequential local fitting, i.e., the entire data of one building are not simultaneously participating in determining the building model. As a consequence, the solution is lack of topological integrity and geometric rigor. Fundamentally different from this traditional approach, we propose a holistic parametric reconstruction method which means taking into consideration the entire point clouds of one building simultaneously. In our work, building models are reconstructed from predefined parametric (roof) primitives. We first use a well-designed deep neural network to segment and identify primitives in the given building point clouds. A holistic optimization strategy is then introduced to simultaneously determine the parameters of a segmented primitive. In the last step, the optimal parameters are used to generate a watertight building model in CityGML format. The airborne LiDAR dataset RoofN3D with predefined roof types is used for our test. It is shown that PointNet++ applied to the entire dataset can achieve an accuracy of 83% for primitive classification. For a subset of 910 buildings in RoofN3D, the holistic approach is then used to determine the parameters of primitives and reconstruct the buildings. The achieved overall quality of reconstruction is 0.08 meters for point-surface-distance or 0.7 times RMSE of the input LiDAR points. This study demonstrates the efficiency and capability of the proposed approach and its potential to handle large scale urban point clouds.

The first release of the AST3-1 Point Source Catalogue from Dome A, Antarctica

The three Antarctic Survey Telescopes (AST3) aim to carry out time domain imaging survey at Dome A, Antarctica. The first of the three telescopes (AST3-1) was successfully deployed on January 2012. AST3-1 is a 500\,mm aperture modified Schmidt telescope with a 680\,mm diameter primary mirror. AST3-1 is equipped with a SDSS $i$ filter and a 10k $\times$ 10k frame transfer CCD camera, reduced to 5k $\times$ 10k by electronic shuttering, resulting in a 4.3 deg$^2$ field-of-view. To verify the capability of AST3-1 for a variety of science goals, extensive commissioning was carried out between March and May 2012. The commissioning included a survey covering 2000 deg$^2$ as well as the entire Large and Small Magellanic Clouds. Frequent repeated images were made of the center of the Large Magellanic Cloud, a selected exoplanet transit field, and fields including some Wolf-Rayet stars. Here we present the data reduction and photometric measurements of the point sources observed by AST3-1. We have achieved a survey depth of 19.3\,mag in 60 s exposures with 5\,mmag precision in the light curves of bright stars. The facility achieves sub-mmag photometric precision under stable survey conditions, approaching its photon noise limit. These results demonstrate that AST3-1 at Dome A is extraordinarily competitive in time-domain astronomy, including both quick searches for faint transients and the detection of tiny transit signals.

GROUND FILTERING LiDAR DATA BASED ON MULTI-SCALE ANALYSIS OF HEIGHT DIFFERENCE THRESHOLD

Abstract. Separating point clouds into ground and non-ground points is a necessary step to generate digital terrain model (DTM) from LiDAR dataset. In this research, a new method based on multi-scale analysis of height difference threshold is proposed for ground filtering of LiDAR data. The proposed method utilizes three windows with different sizes in small, average and large to cover the entire LiDAR point clouds, then with a height difference threshold, point clouds can be separated to ground and non-ground in each local window. Meanwhile, the best threshold values for size of windows are considered based on physical characteristics of the ground surface and size of objects. Also, the minimum of height of object in each window selected as height difference threshold. In order to evaluate the performance of the proposed algorithm, two datasets in rural and urban area were applied. The overall accuracy in rural and urban area was 96.06% and 94.88% respectively. These results of the filtering showed that the proposed method can successfully filters non-ground points from LiDAR point clouds despite of the data area.

Urban building reconstruction from raw LiDAR point data

We present a method for automatic reconstruction of the volumetric structures of urban buildings, directly from raw LiDAR point clouds. Given the large-scale LiDAR data from a group of urban buildings, we take advantage of the “divide-and-conquer” strategy to decompose the entire point clouds into a number of subsets, each of which corresponds to an individual building. For each urban building, we determine its upward direction and partition the corresponding point data into a series of consecutive blocks, achieved by investigating the distributions of feature points of the building along the upward direction. Next, we propose a novel algorithm, Spectral Residual Clustering (SRC), to extract the primitive elements within the contours of blocks from the sectional point set, which is formed by registering the series of consecutive slicing points. Subsequently, we detect the geometric constraints among primitive elements through individual fitting, and perform constrained fitting over all primitive elements to obtain the accurate contour. On this basis, we execute 3D modeling operations, like extrusion, lofting or sweeping, to generate the 3D models of blocks. The final accurate 3D models are generated by applying the union Boolean operations over the block models. We evaluate our reconstruction method on a variety of raw LiDAR scans to verify its robustness and effectiveness.

Three-dimensional object reconstruction using patch significance correspondence

We propose an approach to reconstruct a three-dimensional model from a stream of RGB-D images. Compared with existing methods, our strategy performs well in challenging cases, involving quick movements of the camera and missing depth values. The key idea underlying our approach is to combine registration with patch segmentation based on RGB information. We use these segments and the patch significance correspondence algorithm to transform a global restriction into a number of local restrictions during camera moving. Furthermore, we propose a method that improves the low precision of geometric registration by aligning corresponding patches instead of entire point clouds. In light of this consideration of spatial relations, we also propose a fusion strategy to extract the correct transformations in spite of poor RGB-D sequences. The results of tests on RGB-D benchmark sequences and comparisons with KinectFusion system showed that the proposed approach substantially increases the accuracy of the reconstructed models.

RESEARCH INTO THE COLLIMATION AND HORIZONTAL AXIS ERRORS INFLUENCE ON THE Z+F LASER SCANNER ACCURACY OF VERTICALITY MEASUREMENT

Abstract. Aim of this study was to appoint values of collimation and horizontal axis errors of the laser scanner ZF 5006h owned by Department of Geodesy and Cartography, Warsaw University of Technology, and then to determine the effect of those errors on the results of measurements. An experiment has been performed, involving measurement of the test field , founded in the Main Hall of the Main Building of the Warsaw University of Technology, during which values of instrumental errors of interest were determined. Then, an universal computer program that automates the proposed algorithm and capable of applying corrections to measured target coordinates or even entire point clouds from individual stations, has been developed.

RESEARCH INTO THE COLLIMATION AND HORIZONTAL AXIS ERRORS INFLUENCE ON THE Z+F LASER SCANNER ACCURACY OF VERTICALITY MEASUREMENT

Aim of this study was to appoint values of collimation and horizontal axis errors of the laser scanner ZF 5006h owned by Department of Geodesy and Cartography, Warsaw University of Technology, and then to determine the effect of those errors on the results of measurements. An experiment has been performed, involving measurement of the test field , founded in the Main Hall of the Main Building of the Warsaw University of Technology, during which values of instrumental errors of interest were determined. Then, an universal computer program that automates the proposed algorithm and capable of applying corrections to measured target coordinates or even entire point clouds from individual stations, has been developed.

Observational studies of the formation and evolution of dense cores

AbstractDense cores are the simplest star-forming sites. They represent the end stage of the fragmentation hierarchy that characterizes molecular clouds, and they likely control the efficiency of star formation via their relatively low numbers. Recent dust continuum observations of entire molecular clouds show that dense cores often lie along large-scale filamentary structures, suggesting that the cores form by some type of fragmentation process in an approximately cylindrical geometry. To understand the formation mechanism of cores, additional kinematic information is needed, and this requires observations in molecular-line tracers of both the dense cores and their surrounding cloud material. Here I present some recent efforts to clarify the kinematic structure of core-forming regions in the nearby Taurus molecular cloud. These new observations show that the filamentary structures seen in clouds are often more complex than suggested by the maps of continuum emission, and that they consist of multiple fiber-like components that have different velocities and sonic internal motions. These components likely arise from turbulent fragmentation of the large-scale flows that generate the filamentary structures. While not all these fiber-like components further fragment to form dense cores, a small group of them does so, likely by gravitational instability. This fragmentation produces characteristic chain-like groups of dense cores that further evolve to form stars.

PRACTICAL IMPLEMENTATION OF SEMI-AUTOMATED AS-BUILT BIM CREATION FOR COMPLEX INDOOR ENVIRONMENTS

Abstract. In recent days, for efficient management and operation of existing buildings, the importance of as-built BIM is emphasized in AEC/FM domain. However, fully automated as-built BIM creation is a tough issue since newly-constructed buildings are becoming more complex. To manage this problem, our research group has developed a semi-automated approach, focusing on productive 3D as-built BIM creation for complex indoor environments. In order to test its feasibility for a variety of complex indoor environments, we applied the developed approach to model the ‘Charlotte stairs’ in Lotte World Mall, Korea. The approach includes 4 main phases: data acquisition, data pre-processing, geometric drawing, and as-built BIM creation. In the data acquisition phase, due to its complex structure, we moved the scanner location several times to obtain the entire point clouds of the test site. After which, data pre-processing phase entailing point-cloud registration, noise removal, and coordinate transformation was followed. The 3D geometric drawing was created using the RANSAC-based plane detection and boundary tracing methods. Finally, in order to create a semantically-rich BIM, the geometric drawing was imported into the commercial BIM software. The final as-built BIM confirmed that the feasibility of the proposed approach in the complex indoor environment.

Kinect Depth Data Segmentation Based on Gauss Mixture Model Clustering

Indoor scene understanding based on the depth image data is a cutting-edge issue in the field of three-dimensional computer vision. Taking the layout characteristics of the indoor scenes and more plane features in these scenes into account, this paper presents a depth image segmentation method based on Gauss Mixture Model clustering. First, transform the Kinect depth image data into point cloud which is in the form of discrete three-dimensional point data, and denoise and down-sample the point cloud data; second, calculate the point normal of all points in the entire point cloud, then cluster the entire normal using Gaussian Mixture Model, and finally implement the entire point clouds segmentation by RANSAC algorithm. Experimental results show that the divided regions have obvious boundaries and segmentation quality is above normal, and lay a good foundation for object recognition.

Rotationally Excited H2 in the Magellanic Clouds

AbstractWe have performed a survey study of rotational excited-state H2 Lyman-Werner absorption lines in the entire FUSE Magellanic Clouds Legacy archive. These lines reflect the UV pumping and formation conditions of H2, enabling a more comprehensive study of H2 gas properties, e.g. J-level populations N(J) and b-values (generally indicating the velocity dispersion). Combining with our previous measurements of N(Hi) and N(H2), we derived H2 excitation temperatures, gas volume density n(H), and local UV radiation field strength IUV for each sight line. The results indicate a weaker correlation between n(H) and IUV in Magellanic Clouds than the Galactic sight lines. We also obtained N(H)/E(B − V) ratios from the Spitzer-SAGE and previous CO J = 1 − 0 / Hi 21 cm surveys at sight line locations, using dust modeling and standard line brightness-column density conversion factors. They show a roughly linear correlation with absorption-based N(H)/E(B − V) values, and have a similar scatter (∼0.7 dex) across the LMC and SMC.

Oxygen- and carbon-rich variable red giant populations in the Magellanic Clouds from EROS, OGLE, MACHO, and 2MASS photometry

The carbon-to-oxygen (C/O) ratio of asymptotic giant branch (AGB) stars constitutes an important index of evolutionary and environment/metallicity factor. We develop a method for mass C/O classification of AGBs in photometric surveys without using periods. For this purpose we rely on the slopes in the tracks of individual stars in the colour-magnitude diagram. We demonstrate that our method enables the separation of C-rich and O-rich AGB stars with little confusion. For the Magellanic Clouds we demonstrate that this method works for several photometric surveys and filter combinations. As we rely on no period identification, our results are relatively insensitive to the phase coverage, aliasing, and time-sampling problems that plague period analyses. For a subsample of our stars, we verify our C/O classification against published C/O catalogues. With our method we are able to produce C/O maps of the entire Magellanic Clouds. Our purely photometric method for classification of C- and O-rich AGBs constitutes a method of choice for large, near-infrared photometric surveys. Because our method depends on the slope of colour-magnitude variation but not on magnitude zero point, it remains applicable to objects with unknown distances.

Searching for chameleon-like scalar fields with the ammonia method

(Abridged) In our previous work we found a statistically significant offset Delta V = 27 m/s between the radial velocities of the HC3N(2-1) and NH3(1,1) transitions observed in molecular cores from the Milky Way. This may indicate that the electron-to-proton mass ratio, mu = m_e/m_p, increases by 3x10^{-8} when measured under interstellar conditions with matter densities of more than 10 orders of magnitude lower as compared with laboratory (terrestrial) environments. We now map four molecular cores L1498, L1512, L1517, and L1400K selected from our previous sample in order to estimate systematic effects in Delta V due to possible velocity gradients. We find that in two cores L1498 and L1512 the NH3(1,1) and HC3N(2-1) transitions closely trace the same material and show an offset of Delta V = 26.9 +/- 1.2_stat +/- 3.0_sys m/s throughout the entire clouds. The measured velocity offset, being expressed in terms of Delta mu = (mu_obs - mu_lab)/mu_lab, gives Delta mu = (26 +/- 1_stat +/- 3_sys)x10^{-9}.

THE MASS-SIZE RELATION FROM CLOUDS TO CORES. I. A NEW PROBE OF STRUCTURE IN MOLECULAR CLOUDS

We use a new contour-based map analysis technique to measure the mass and size of molecular cloud fragments continuously over a wide range of spatial scales (0.05 < r / pc < 10), i.e., from the scale of dense cores to those of entire clouds. The present paper presents the method via a detailed exploration of the Perseus Molecular Cloud. Dust extinction and emission data are combined to yield reliable scale-dependent measurements of mass. This scale-independent analysis approach is useful for several reasons. First, it provides a more comprehensive characterization of a map (i.e., not biased towards a particular spatial scale). Such a lack of bias is extremely useful for the joint analysis of many data sets taken with different spatial resolution. This includes comparisons between different cloud complexes. Second, the multi-scale mass-size data constitutes a unique resource to derive slopes of mass-size laws (via power-law fits). Such slopes provide singular constraints on large-scale density gradients in clouds.