Three-dimensional Point Cloud Data Research Articles

High-volume point cloud data simplification based on decomposed graph filtering

Recent studies on three-dimensional (3D) point cloud data (PCD) simplification have played significant roles in computer-aided models for alleviating computational and storage burden. However, existing simplification methods are not suitable for the high-volume PCD with billions of point number, especially in construction industry. In this paper, a decomposed simplification method is developed to handle the high-volume buildings' PCD. Based on the divide-and-conquer philosophy, the new decomposed approach effectively reduces the memory usage. In the proposed approach, PCD is divided into several subsets according to the relationship of natural neighbor, and decomposed graph filtering with adaptive resampling rate is designed. It can be proved that the decomposed simplification results have no variance in the comparison with the optimal simplification of entire PCD. Verification experiments are conducted on different sizes of PCD, which indicate the effectiveness and feasibility of the proposed approach.

WEAKLY SUPERVISED PSEUDO-LABEL ASSISTED LEARNING FOR ALS POINT CLOUD SEMANTIC SEGMENTATION

Abstract. Competitive point cloud semantic segmentation results usually rely on a large amount of labeled data. However, data annotation is a time-consuming and labor-intensive task, particularly for three-dimensional point cloud data. Thus, obtaining accurate results with limited ground truth as training data is considerably important. As a simple and effective method, pseudo labels can use information from unlabeled data for training neural networks. In this study, we propose a pseudo-label-assisted point cloud segmentation method with very few sparsely sampled labels that are normally randomly selected for each class. An adaptive thresholding strategy was proposed to generate a pseudo-label based on the prediction probability. Pseudo-label learning is an iterative process, and pseudo labels were updated solely on ground-truth weak labels as the model converged to improve the training efficiency. Experiments using the ISPRS 3D sematic labeling benchmark dataset indicated that our proposed method achieved an equally competitive result compared to that using a full supervision scheme with only up to 2‰ of labeled points from the original training set, with an overall accuracy of 83.7% and an average F1 score of 70.2%.

The effectiveness of a method that uses stabilized cameras and photogrammetry to survey the size and distribution of individual trees in a mangrove forest

ABSTRACT For the conservation of mangrove forests, management plans should be prepared after evaluating the decreases of mangrove forests explicitly. Although information regarding stand development in mangrove forests is indispensable, there is a lack of it. We reasoned that gathering three-dimensional point cloud data using stabilized cameras and photogrammetry (developed together with unmanned aerial vehicles) could be used to conduct a survey of the sizes and spatial distribution of individual trees in mangrove forests. We surveyed two mangrove forests growing at the mouth of the Okukubi and Ōura rivers on Okinawa Island. The forest floor was photographed using a GoPro Hero 4 attached to a camera stabilizer on about 2- m- long extension pole. The resulting images were processed with a photogrammetry software to create an orthomosaic image. The trunks of trees and seedlings in all plots were visually interpreted. The total number of the seedlings and trees in each plot was counted, and the density in each plot was calculated. The diameter at breast height (DBH) of the trees was calculated from the area of the polygons drawn around the trunks manually. There was a strong correlation between the interpreted and measured DBH. We confirmed the effectiveness of a method that uses stabilized cameras and photogrammetry to survey the size and distribution of individual trees in a mangrove forest. The information obtained by present study is considered to be an essential forest inventory of mangrove forests.

An Alternative Method for Long-Term Monitoring of Thai Historic Pagodas Based on Terrestrial Laser Scanning Data: A Case Study of Wat Krachee in Ayutthaya

This article presents an alternative method for the long-term monitoring of heritage pagodas in Thailand. In this method, terrestrial laser scanning (TLS) is used in combination with permanent survey markers. The Wat (temple) Krachee in the Ayutthaya Province of Thailand was chosen as a case study. This temple has several fantastic elements, including an inverted bell-shaped pagoda, two intertwined trees growing within it, and a chamber inside the pagoda. The preservation team working on the pagoda encountered a challenging problem and faced a decision to trim or not to trim the tree since it has a long-term effect on the pagoda’s structural stability. A high-accuracy terrestrial laser scanner was used to collect three-dimensional point cloud data. Permanent survey markers were constructed in 2018 to be used in long-term monitoring. The 3D surveying of the temple and the monitoring of the pagoda were carried out in five sessions during a period ending in 2020. A point cloud data analysis was performed to obtain the current dimensions, a displacement analysis, and the pagoda leaning angle. The results revealed that the terrestrial laser scanner is a high-performance piece of equipment offering efficient evaluation and long-term monitoring. However, in this study, permanent survey markers were also required as a benchmark for constraining each monitoring session. The 3D point cloud models could be matched with the assumption model elements to evaluate the damaged shape and to determine the original form. The significant elements of an inverted bell-shaped pagoda were investigated. Trimming the tree was found to cause the leaning angle of the pagoda to decrease. An equation was developed for predicting the leaning angle of the Wat Krachee pagoda for preservation and restoration planning in the future. From the results of this study, it is recommended that periodic monitoring should continue in order to preserve Thai pagodas in their original forms.

Multi-spacecraft collaborative attitude determination of space tumbling target with experimental verification

The precise attitude determination of a tumbling non-cooperative space target is challenging. This paper proposed an attitude determination approach based on multi-spacecraft collaborative observation. The three-dimensional (3-D) point cloud data of the target is constructed in terms of the collaborative observation data. By virtue of the target features and point cloud matching between two adjacent frames, the attitude change information of the target is derived, and based on this, the attitude information of the tumbling target is estimated via Extended Kalman Filter (EKF). Numerical simulations and ground micro-gravity experiments are carried out to demonstrate the performance of the proposed collaborative observation method. The results show that the presented approach can successfully estimate the attitude of the tumbling target with no prior information, and avoid the presence of occlusion or data missing during the observation process.

Fragmentation calculation method for blast muck piles in open-pit copper mines based on three-dimensional laser point cloud data

The current calculation methods for the blast fragmentation of muck piles (BFMP) do not consider three-dimensional (3D) feature information, such as concavity and convexity. Thus, their calculation results are inaccurate. In this work, 3D laser scanning technology with high precision, high efficiency, and real-time digitization is used to calculate fragmentation of blast muck piles in open-pit copper mines. The voxel cloud connectivity segmentation (VCCS) algorithm and locally convex connected patches (LCCP) algorithm are used to calculate BFMP from 3D laser point cloud data (3D LPCD). Meanwhile, an improved VCCS algorithm based on discrete features is proposed to solve the segmentation problem of 3D LPCD of small blast muck piles, which cannot be effectively segmented and thus greatly influence the 3D LPCD of large blast muck piles. An improved LCCP algorithm based on plane fitting solves the incorrect segmentation of the surface of large fragments into small pieces of fragments. Dexing Copper Mine, the largest open-pit copper mine in China, is taken as the research object. Results show that the accuracy of the calculation results is about 80% when the fragment sizes are 0.1–0.5 m. Moreover, almost all BFMP can be correctly calculated when the fragment size exceeds 0.5 m.

An assessment of conventional and drone-based measurements for tree attributes in timber volume estimation: A case study on stone pine plantation

The use of unmanned aerial vehicles (UAVs) as a useful tool in forestry studies is increasing rapidly. The level of accuracy in UAV-supported measurement data has been increasing in the field of forestry. Recently, biophysical and morphological tree information has been calculated using the three-dimensional (3D) point cloud data. The accuracy of studies on different types of trees (i.e. coniferous and deciduous) may vary depending on the selected instruments and methods. Within the scope of the present study, height (H) and crown projection area (CPA) of 105 stone pines (Pinus pinea, L.) measured using both conventional methods and UAV-based Structure-from-Motion (SfM) derived 3D dense point cloud were evaluated. Tree volumes in the study area were calculated using the allometric formulas generated based on the linear diameter model produced with the field measurements. For field measurements and UAV based data, the tree volumes were calculated using the single- and double-entry over-bark stem volume equations, and their numerical comparisons were conducted. To search for the differences, RMSE (Root mean square error), RMSE% (Root mean square percentage error), MAE (Mean absolute error) and MAE% (Mean absolute percentage error) values were taken into consideration. According to the results of the paired t-test, it was revealed that there were no significant differences between the field- and SfM- measurements based methods. Considering the average values instead of individual (single) trees in the information obtained using the 3D point cloud in such stands gave more accurate results. It was found that forest parameters at plot levels in stands could be quickly revealed by UAV photogrammetry. In addition, these data can be evaluated as a metric measurement technique for sustainable and precise operational planning in forest lands.

Automatic recognition and 3D modeling of the neck-shoulder human shape based on 2D images

PurposeThis study focused on how to realize automatic recognition of young women's neck-shoulder shape based on the front and side images.Design/methodology/approachThe reverse engineering software was used to measure the body sizes of the neck-shoulder part based on the young women's three-dimensional (3D) point cloud data, and the important parameters closely related to the neck-shoulder shape were determined. The neck-shoulder shape of the subjects was classified to establish the classification rules. Then, based on the front and side images, the human body contour was extracted by Matlab, and the data required for neck-shoulder shape classification were obtained by identifying the feature points.FindingsThrough the cluster analysis based on the shoulder angle, back angle, shoulder depth/width ratio and armpit depth/width ratio, young women's neck-shoulder shape was divided into four categories, namely round wide shoulder, flat narrow shoulder, round drop shoulder and hunchback flat shoulder. The neck-shoulder shape could be automatically recognized based on the established classification rules and two-dimensional (2D) body measurement method, with an accuracy rate of 90%. The neck-shoulder shape automatic recognition system constructed based on this method is effective.Originality/valueThis study proposed a simple neck-shoulder automatic recognition method based on the 2D body images. This approach can be extended to other group of human body or other parts of the body.

Fully automated methodology for the delineation of railway lanes and the generation of IFC alignment models using 3D point cloud data

This work presents a fully automated methodology that, in a first step, is able to reliably extract and delineate the position and geometry of the rails from three-dimensional (3D) point cloud data of railway infrastructure by sequentially applying heuristic-based point cloud processing steps, namely railway track segmentation, rough rail estimation, and rail extraction. Then, that information is used to generate the alignment of the surveyed railway lane following the requirements of the Industry Foundation Classes (IFC) Alignment standard. Finally, the proposed method exports an IFC-compliant file that describes the position of the rails with respect to the railway lane alignment. This method has been applied to a 90-km long railway lane, and validated on two one-kilometer subsections of the case study data, obtaining an average rail delineation error of less than 3 cm. Furthermore, the track gauge was measured using the rail alignment data, obtaining errors of a similar magnitude.

Three-Dimensional Extended Object Tracking and Shape Learning Using Gaussian Processes

In this study, we investigate the problem of tracking objects with unknown shapes using three-dimensional (3D) point cloud data. We propose a Gaussian process-based model to jointly estimate object kinematics, including position, orientation and velocities, together with the shape of the object for online and offline applications. We describe the unknown shape by a radial function in 3D, and induce a correlation structure via a Gaussian process. Furthermore, we propose an efficient algorithm to reduce the computational complexity of working with 3D data. This is accomplished by casting the tracking problem into projection planes which are attached to the object's local frame. The resulting algorithms can process 3D point cloud data and accomplish tracking of a dynamic object. Furthermore, they provide analytical expressions for the representation of the object shape in 3D, together with confidence intervals. The confidence intervals, which quantify the uncertainty in the shape estimate, can later be used for solving the gating and association problems inherent in object tracking. The performance of the methods is demonstrated both on simulated and real data. The results are compared with an existing random matrix model, which is commonly used for extended object tracking in the literature.

High-Precision Method for Estimating the Three-Dimensional Green Quantity of an Urban Forest

Forests are an important component of urban ecosystems that offer a variety of ecological services; however, there are few reports on the application of high-precision three-dimensional (3D) laser-scanned point cloud data to quantify the 3D green volume. To quantitatively study the 3D green quantity of urban forests in Xi'an, China, aerial remote sensing and high-precision 3D laser-scanned point cloud data were used to build an estimation model using a differential method. The results show that the 3D green quantity model has a high estimation accuracy of 88.07%, with an R value of 0.89, and can be used as an inverse model for estimating the 3D standing wood volume. The method can be automated for data processing and has a high measurement efficiency, making it applicable to real situations and significantly improving the measurement accuracy of single-tree 3D green quantification. The total 3D green quantity in the main urban area of Xi’an is 37,475,294.7 m3, where areas containing the greatest green volume follow the order scenic sites > cultural and educational areas > traffic routes > residential areas > industrial areas. This method can be used to improve urban design, policy making, and healthy urban forests.

Design of supercontinuum laser hyperspectral light detection and ranging (LiDAR) (SCLaHS LiDAR)

ABSTRACT Traditional Light Detection and Rangings (LiDARs) can quickly collect high-accuracy of three-dimensional (3D) point cloud data at a designated wavelength (i.e., cannot obtain hyperspectral data), while the passive hyperspectral imager can collect rich spectral data of ground objects, but are lack of 3D spatial data. This paper presents one innovative study on the design of airborne-oriented supercontinuum laser hyperspectral (SCLaHS) LiDAR with 50 bands covering 400 nm to 900 nm at a spectral resolution of 10 nm and ground sampling distance (GSD) of 0.5 m. The major innovations include (1) development of the high-power narrow-pulse supercontinuum laser source covering 400 nm to 900 nm with 50 bands using multi-core microstructure fibre, all-polarization maintaining fibre and ultra-long cavity structure, (2) a miniaturized aberration correction holographic concave grating spectroscopic and streak tube technique are developed for 50 bands laser echoes detection at high spectral-spatial-temporal resolution and dynamic airborne platform, and (3) the algorithm theoretic basis for SCLaHS LiDAR point cloud data 3D geodetic coordination calculation, including in-flight airborne calibration algorithm. The initial experimental results demonstrated that the designed SCLaHS LiDAR is doable, and a prototype of the (SCLaHS) LiDAR intends to be implemented.

Asphalt pavement friction coefficient prediction method based on genetic-algorithm-improved neural network (GAI-NN) model

To overcome the limitations of pavement skid resistance prediction using the friction coefficient, a genetic-algorithm-improved neural network (GAI-NN) was developed in this study. First, three-dimensional (3D) point-cloud data of an asphalt pavement surface were obtained using a smart sensor (Gocator 3110). The friction coefficient of the pavement was then obtained using a pendulum friction tester. The 3D point-cloud dataset was then analyzed to recover the missing data and perform denoising. In particular, these data were filled using cubic spline interpolation. Parameters for texture characterization were defined, and methods for computing the parameters were developed. Finally, the GAI-NN model was developed by modifying the weights and thresholds. The test results indicated that using pavement surface texture 3D data, the GAI-NN was capable of predicting the pavement friction coefficient with sufficient accuracy, with an error of 12.1%.

Management Point Cloud Data based on GeoSOT-3D Subdivision Model

Owing to the rapid development of earth observation technology, the volume of point cloud dataset is growing rapidly. The efficient management and retrieval method is very important for effective use of the large point cloud datasets. Current methods for point cloud data management are generally based on file which makes it difficult to describe internal structure of the modelling objects, while current indexing is generally based on coordinates which is quite time-consuming for large volume cloud point data. To solve these problems, we proposed a method to manage and index point cloud data with a global grid model GeoSOT. The subdivision model of three-dimensional point cloud data was defined to identify the global position of point clouds and internal points with GeoSOT-3D grids. The management and indexing of point clouds were implemented based on the GeoSOT-3D grid codes integrated with the metadata table. A 3D prototype platform for showing the method was developed based on the Cesium. Experimental results show that our method is able to manage, query, visualize massive point clouds as well as other data sources in a unified global position framework.

3DPhenoFish: Application for two- and three-dimensional fish morphological phenotype extraction from point cloud analysis.

Fish morphological phenotypes are important resources in artificial breeding, functional gene mapping, and population-based studies in aquaculture and ecology. Traditional morphological measurement of phenotypes is rather expensive in terms of time and labor. More importantly, manual measurement is highly dependent on operational experience, which can lead to subjective phenotyping results. Here, we developed 3DPhenoFish software to extract fish morphological phenotypes from three-dimensional (3D) point cloud data. Algorithms for background elimination, coordinate normalization, image segmentation, key point recognition, and phenotype extraction were developed and integrated into an intuitive user interface. Furthermore, 18 key points and traditional 2D morphological traits, along with 3D phenotypes, including area and volume, can be automatically obtained in a visualized manner. Intuitive fine-tuning of key points and customized definitions of phenotypes are also allowed in the software. Using 3DPhenoFish, we performed high-throughput phenotyping for four endemic Schizothoracinae species, including Schizopygopsis younghusbandi, Oxygymnocypris stewartii, Ptychobarbus dipogon, and Schizothorax oconnori. Results indicated that the morphological phenotypes from 3DPhenoFish exhibited high linear correlation (>0.94) with manual measurements and offered informative traits to discriminate samples of different species and even for different populations of the same species. In summary, we developed an efficient, accurate, and customizable tool, 3DPhenoFish, to extract morphological phenotypes from point cloud data, which should help overcome traditional challenges in manual measurements. 3DPhenoFish can be used for research on morphological phenotypes in fish, including functional gene mapping, artificial selection, and conservation studies. 3DPhenoFish is an open-source software and can be downloaded for free at https://github.com/lyh24k/3DPhenoFish/tree/master.

Research on point cloud generation algorithm of virtual depth camera

In this paper, based on the example of insulator photography with real sense depth camera, we use OpenGL to design a virtual camera to photograph the three-dimensional model of insulator STL, extract the surface data information of the three-dimensional model of insulator STL, and then obtain the three-dimensional point cloud data of insulator. Finally, random noise is added to the point cloud data obtained by the virtual camera to simulate the real situation of the insulator captured by the real sense depth camera. The experimental results show that the virtual depth camera can replace the solid camera to collect the point cloud data of the STL 3D model of the object.

Research on point cloud generation algorithm of virtual depth camera

In this paper, based on the example of insulator photography with real sense depth camera, we use OpenGL to design a virtual camera to photograph the three-dimensional model of insulator STL, extract the surface data information of the three-dimensional model of insulator STL, and then obtain the three-dimensional point cloud data of insulator. Finally, random noise is added to the point cloud data obtained by the virtual camera to simulate the real situation of the insulator captured by the real sense depth camera. The experimental results show that the virtual depth camera can replace the solid camera to collect the point cloud data of the STL 3D model of the object.

Simulation of Skin Stretching around the Forehead Wrinkles in Rhytidectomy

Skin stretching around the forehead wrinkles is an important method in rhytidectomy. Proper parameters are required to evaluate the surgical effect. In this paper, a simulation method was proposed to obtain the parameters. Three-dimensional point cloud data with a resolution of 50 μm were employed. First, a smooth supporting contour under the wrinkled forehead was generated via b-spline interpolation and extrapolation to constrain the deformation of the wrinkled zone. Then, based on the vector formed intrinsic finite element (VFIFE) algorithm, the simulation was implemented in Matlab for the deformation of wrinkled forehead skin in the stretching process. Finally, the stress distribution and the residual wrinkles of forehead skin were employed to evaluate the surgical effect. Although the residual wrinkles are similar when forehead wrinkles are finitely stretched, their stress distribution changes greatly. This indicates that the stress distribution in the skin is effective to evaluate the surgical effect, and the forehead wrinkles are easily to be overstretched, which may lead to potential skin injuries. The simulation method can predict stress distribution and residual wrinkles after forehead wrinkle stretching surgery, which can be potentially used to control the surgical process and further reduce risks of skin injury.

A TERRESTRIAL LASER SCANNING MEASUREMENT STATION TO MONITOR LONG-TERM STRUCTURAL DYNAMICS IN A BOREAL FOREST

Abstract. This work presents a data acquisition framework and the technical details of a permanent terrestrial laser scanning (TLS) measurement station for high spatial and temporal resolution forest observation that was developed in the Finnish Geospatial Research Institute. The TLS measurement station was established to provide hyper-temporal time series of three-dimensional point cloud data for long term monitoring of a boreal forest. Time series data acquisition framework consists of regular 14-minute scans performed by a RIEGL VZ-2000i laser scanner in every 30 minutes, resulting in the collection of 48 scans per day. The entire framework includes the setting up of the laser scanner, the initialization of daily project, the scanning data acquisition over a preset time window, the storage management of the collected data at a local measurement computer, and the transfer of data from the measurement computer to network-attached storage (NAS) for further data processes. The operability of the proposed TLS measurement station was first piloted at a test area of about 32,500 m2 in Southern Finland (60°09'N, 24°32'E). A set of several long monitoring experiments were performed over the whole growing season from the beginning of April to the end of October in 2019. As preliminary results, the time series outputs have captured detailed information on the phenological changes in the test site with sub-centimetre accuracy. For instance, it was possible to visualize plant dynamics phenomena, such as the sprouting of leaves in spring and their falling in autumn.

Swept Source Lidar: simultaneous FMCW ranging and nonmechanical beam steering with a wideband swept source.

Light detection and ranging (lidar) has long been used in various applications. Solid-state beam steering mechanisms are needed for robust lidar systems. Here we propose and demonstrate a lidar scheme called "Swept Source Lidar" that allows us to perform frequency-modulated continuous-wave (FMCW) ranging and nonmechanical beam steering simultaneously. Wavelength dispersive elements provide angular beam steering, while a laser frequency is continuously swept by a wideband swept source over its whole tuning bandwidth. Employing a tunable vertical-cavity surface-emitting laser and a 1-axis mechanical beam scanner, three-dimensional point cloud data has been obtained. Swept Source Lidar systems can be flexibly combined with various beam steering elements to realize full solid-state FMCW lidar systems.