A Random Fusion of Mix 3D and Polar Mix to Improve Semantic Segmentation Performance in 3D Lidar Point Cloud

This paper focuses on the effective utilization of data augmentation techniques for 3Dlidar point clouds to enhance the performance of neural network models.These point clouds,which represent spatial information through a collection of 3D coordinates,have found wide-ranging applications.Data augmentation has emerged as a potent solution to the challenges posed by limited labeled data and the need to enhance model generalization capabilities.Much of the existing research is devoted to crafting novel data augmentation methods specifically for 3D lidar point clouds.However,there has been a lack of focus on making the most of the numerous existing augmentation techniques.Addressing this deficiency,this research investigates the possibility of combining two fundamental data augmentation strategies.The paper introduces PolarMix andMix3D,two commonly employed augmentation techniques,and presents a new approach,named RandomFusion.Instead of using a fixed or predetermined combination of augmentation methods,RandomFusion randomly chooses one method from a pool of options for each instance or sample.This innovative data augmentation technique randomly augments each point in the point cloud with either PolarMix or Mix3D.The crux of this strategy is the random choice between PolarMix and Mix3Dfor the augmentation of each point within the point cloud data set.The results of the experiments conducted validate the efficacy of the RandomFusion strategy in enhancing the performance of neural network models for 3D lidar point cloud semantic segmentation tasks.This is achieved without compromising computational efficiency.By examining the potential of merging different augmentation techniques,the research contributes significantly to a more comprehensive understanding of how to utilize existing augmentation methods for 3D lidar point clouds.RandomFusion data augmentation technique offers a simple yet effective method to leverage the diversity of augmentation techniques and boost the robustness of models.The insights gained from this research can pave the way for future work aimed at developing more advanced and efficient data augmentation strategies for 3D lidar point cloud analysis.

SGT-Net: A Transformer-Based Stratified Graph Convolutional Network for 3D Point Cloud Semantic Segmentation

In recent years,semantic segmentation on 3D point cloud data has attracted much attention.Unlike 2D images where pixels distribute regularly in the image domain,3D point clouds in non-Euclidean space are irregular and inherently sparse.Therefore,it is very difficult to extract long-range contexts and effectively aggregate local features for semantic segmentation in 3D point cloud space.Most current methods either focus on local feature aggregation or long-range context dependency,but fail to directly establish a global-local feature extractor to complete the point cloud semantic segmentation tasks.In this paper,we propose a Transformer-based stratified graph convolutional network(SGT-Net),which enlarges the effective receptive field and builds direct long-range dependency.Specifically,we first propose a novel dense-sparse sampling strategy that provides dense local vertices and sparse long-distance vertices for subsequent graph convolutional network(GCN).Secondly,we propose a multi-key self-attention mechanism based on the Transformer to further weight augmentation for crucial neighboring relationships and enlarge the effective receptive field.In addition,to further improve the efficiency of the network,we propose a similarity measurement module to determine whether the neighborhood near the center point is effective.We demonstrate the validity and superiority of our method on the S3DIS and ShapeNet datasets.Through ablation experiments and segmentation visualization,we verify that the SGT model can improve the performance of the point cloud semantic segmentation.

Three-dimensional(3D)parametric measurements of individual gravels in the Gobi region using point cloud technique

Gobi spans a large area of China,surpassing the combined expanse of mobile dunes and semi-fixed dunes.Its presence significantly influences the movement of sand and dust.However,the complex origins and diverse materials constituting the Gobi result in notable differences in saltation processes across various Gobi surfaces.It is challenging to describe these processes according to a uniform morphology.Therefore,it becomes imperative to articulate surface characteristics through parameters such as the three-dimensional(3D)size and shape of gravel.Collecting morphology information for Gobi gravels is essential for studying its genesis and sand saltation.To enhance the efficiency and information yield of gravel parameter measurements,this study conducted field experiments in the Gobi region across Dunhuang City,Guazhou County,and Yumen City(administrated by Jiuquan City),Gansu Province,China in March 2023.A research framework and methodology for measuring 3D parameters of gravel using point cloud were developed,alongside improved calculation formulas for 3D parameters including gravel grain size,volume,flatness,roundness,sphericity,and equivalent grain size.Leveraging multi-view geometry technology for 3D reconstruction allowed for establishing an optimal data acquisition scheme characterized by high point cloud reconstruction efficiency and clear quality.Additionally,the proposed methodology incorporated point cloud clustering,segmentation,and filtering techniques to isolate individual gravel point clouds.Advanced point cloud algorithms,including the Oriented Bounding Box(OBB),point cloud slicing method,and point cloud triangulation,were then deployed to calculate the 3D parameters of individual gravels.These systematic processes allow precise and detailed characterization of individual gravels.For gravel grain size and volume,the correlation coefficients between point cloud and manual measurements all exceeded 0.9000,confirming the feasibility of the proposed methodology for measuring 3D parameters of individual gravels.The proposed workflow yields accurate calculations of relevant parameters for Gobi gravels,providing essential data support for subsequent studies on Gobi environments.

A LiDAR Point Clouds Dataset of Ships in a Maritime Environment

For the first time, this article introduces a LiDAR Point Clouds Dataset of Ships composed of both collected and simulated data to address the scarcity of LiDAR data in maritime applications. The collected data are acquired using specialized maritime LiDAR sensors in both inland waterways and wide-open ocean environments. The simulated data is generated by placing a ship in the LiDAR coordinate system and scanning it with a redeveloped Blensor that emulates the operation of a LiDAR sensor equipped with various laser beams. Furthermore,we also render point clouds for foggy and rainy weather conditions. To describe a realistic shipping environment, a dynamic tail wave is modeled by iterating the wave elevation of each point in a time series. Finally, networks serving small objects are migrated to ship applications by feeding our dataset. The positive effect of simulated data is described in object detection experiments, and the negative impact of tail waves as noise is verified in single-object tracking experiments. The Dataset is available at https://github.com/zqy411470859/ship_dataset.

Building Facade Point Clouds Segmentation Based on Optimal Dual-Scale Feature Descriptors

To address the current issues of inaccurate segmentation and the limited applicability of segmentation methods for building facades in point clouds, we propose a facade segmentation algorithm based on optimal dual-scale feature descriptors. First, we select the optimal dual-scale descriptors from a range of feature descriptors. Next, we segment the facade according to the threshold value of the chosen optimal dual-scale descriptors. Finally, we use RANSAC (Random Sample Consensus) to fit the segmented surface and optimize the fitting result. Experimental results show that, compared to commonly used facade segmentation algorithms, the proposed method yields more accurate segmentation results, providing a robust data foundation for subsequent 3D model reconstruction of buildings.

Building 3D CityGML models of mining industrial structures using integrated UAV and TLS point clouds

Mining industrial areas with anthropogenic engineering structures are one of the most distinctive features of the real world.3D models of the real world have been increasingly popular with numerous applications,such as digital twins and smart factory management.In this study,3D models of mining engineering structures were built based on the CityGML standard.For collecting spatial data,the two most popular geospatial technologies,namely UAV-SfM and TLS were employed.The accuracy of the UAV survey was at the centimeter level,and it satisfied the absolute positional accuracy requirement of creat-ing all levels of detail(LoD)according to the CityGML standard.Therefore,the UAV-SfM point cloud dataset was used to build LoD 2 models.In addition,the comparison between the UAV-SfM and TLS sub-clouds of facades and roofs indicates that the UAV-SfM and TLS point clouds of these objects are highly consistent,therefore,point clouds with a higher level of detail and accuracy provided by the integration of UAV-SfM and TLS were used to build LoD 3 models.The resulting 3D CityGML models include 39 buildings at LoD 2,and two mine shafts with hoistrooms,headframes,and sheave wheels at LoD3.

Enhanced 3D Point Cloud Reconstruction for Light Field Microscopy Using U-Net-Based Convolutional Neural Networks

This article describes a novel approach for enhancing the three-dimensional(3D)point cloud reconstruction for light field microscopy(LFM)using U-net architecture-based fully convolutional neural network(CNN).Since the directional view of the LFM is limited,noise and artifacts make it difficult to reconstruct the exact shape of 3D point clouds.The existing methods suffer from these problems due to the self-occlusion of the model.This manuscript proposes a deep fusion learning(DL)method that combines a 3D CNN with a U-Net-based model as a feature extractor.The sub-aperture images obtained from the light field microscopy are aligned to form a light field data cube for preprocessing.A multi-stream 3D CNNs and U-net architecture are applied to obtain the depth feature fromthe directional sub-aperture LF data cube.For the enhancement of the depthmap,dual iteration-based weighted median filtering(WMF)is used to reduce surface noise and enhance the accuracy of the reconstruction.Generating a 3D point cloud involves combining two key elements:the enhanced depth map and the central view of the light field image.The proposed method is validated using synthesized Heidelberg Collaboratory for Image Processing(HCI)and real-world LFM datasets.The results are compared with different state-of-the-art methods.The structural similarity index(SSIM)gain for boxes,cotton,pillow,and pens are 0.9760,0.9806,0.9940,and 0.9907,respectively.Moreover,the discrete entropy(DE)value for LFM depth maps exhibited better performance than other existing methods.

Point Cloud Processing Methods for 3D Point Cloud Detection Tasks

Light detection and ranging(LiDAR)sensors play a vital role in acquiring 3D point cloud data and extracting valuable information about objects for tasks such as autonomous driving,robotics,and virtual reality(VR).However,the sparse and disordered nature of the 3D point cloud poses significant challenges to feature extraction.Overcoming limitations is critical for 3D point cloud processing.3D point cloud object detection is a very challenging and crucial task,in which point cloud processing and feature extraction methods play a crucial role and have a significant impact on subsequent object detection performance.In this overview of outstanding work in object detection from the 3D point cloud,we specifically focus on summarizing methods employed in 3D point cloud processing.We introduce the way point clouds are processed in classical 3D object detection algorithms,and their improvements to solve the problems existing in point cloud processing.Different voxelization methods and point cloud sampling strategies will influence the extracted features,thereby impacting the final detection performance.

Indoor Space Modeling and Parametric Component Construction Based on 3D Laser Point Cloud Data

In order to enhance modeling efficiency and accuracy,we utilized 3D laser point cloud data for indoor space modeling.Point cloud data was obtained with a 3D laser scanner and optimized with Autodesk Recap and Revit software to extract geometric information about the indoor environment.Furthermore,we proposed a method for constructing indoor elements based on parametric components.The research outcomes of this paper will offer new methods and tools for indoor space modeling and design.The approach of indoor space modeling based on 3D laser point cloud data and parametric component construction can enhance modeling efficiency and accuracy,providing architects,interior designers,and decorators with a better working platform and design reference.

Aggregate Point Cloud Geometric Features for Processing

As 3D acquisition technology develops and 3D sensors become increasingly affordable,large quantities of 3D point cloud data are emerging.How to effectively learn and extract the geometric features from these point clouds has become an urgent problem to be solved.The point cloud geometric information is hidden in disordered,unstructured points,making point cloud analysis a very challenging problem.To address this problem,we propose a novel network framework,called Tree Graph Network(TGNet),which can sample,group,and aggregate local geometric features.Specifically,we construct a Tree Graph by explicit rules,which consists of curves extending in all directions in point cloud feature space,and then aggregate the features of the graph through a cross-attention mechanism.In this way,we incorporate more point cloud geometric structure information into the representation of local geometric features,which makes our network perform better.Our model performs well on several basic point clouds processing tasks such as classification,segmentation,and normal estimation,demonstrating the effectiveness and superiority of our network.Furthermore,we provide ablation experiments and visualizations to better understand our network.

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.

基于知识蒸馏和定位引导的Pointpillars点云检测网络

激光雷达数据由于其几何特性,被广泛应用于三维目标检测任务中。由于点云数据的稀疏性和不规则性,难以实现特征提取的质量和推理速度间的平衡。本文提出一种基于体柱特征编码的三维目标检测算法,以Pointpillars网络为基础,设计Teacher-Student模型框架对回归框尺度进行蒸馏,增加蒸馏损失,优化训练网络模型,提升特征提取的质量。为进一步提高模型检测效果,设计定位引导分类项,增加分类预测和回归预测之间的相关性,提高物体识别准确率。本网络所做改进没有引入额外的网络嵌入。算法在KITTI数据集上的实验结果表明,相比于基准网络,在三维模式下的平均精度值从60.65%提升到了64.69%,鸟瞰图模式下的平均精度值从67.74%提升到70.24%。模型推理速度为45 FPS,在提升检测精度的同时满足了实时性要求。

A Novel Airborne 3D Laser Point Cloud Hole Repair Algorithm Considering Topographic Features

Hole repair processing is an important part of point cloud data processing in airborne 3-dimensional(3D)laser scanning technology.Due to the fragmentation and irregularity of the surface morphology,when applying the 3D laser scanning technology to mountain mapping,the conventional mathematical cloud-based point cloud hole repair method is not ideal in practical applications.In order to solve this problem,we propose to repair the valley and ridge line first,and then repair the point cloud hole.The main technical steps of the method include the following points:First,the valley and ridge feature lines are extracted by the GIS slope analysis method;Then,the valley and ridge line missing from the hole are repaired by the mathematical interpolation method,and the repaired results are edited and inserted to the original point cloud;Finally,the traditional repair method is used to repair the point cloud hole whose valley line and ridge line have been repaired.Three experiments were designed and implemented in the east bank of the Xiaobaini River to test the performance of the proposed method.The results showed that compared with the direct point cloud hole repair method in Geomagic Studio software,the average repair accuracy of the proposed method,in the 16 m buffer zone of valley line and ridge line,is increased from 56.31 cm to 31.49 cm.The repair performance is significantly improved.

Hand Gesture Recognition Using Appearance Features Based on 3D Point Cloud

This paper presents a method for hand gesture recognition based on 3D point cloud. Digital image processing technology is used in this research. Based on the 3D point from depth camera, the system firstly extracts some raw data of the hand. After the data segmentation and preprocessing, three kinds of appearance features are extracted, including the number of stretched fingers, the angles between fingers and the gesture region’s area distribution feature. Based on these features, the system implements the identification of the gestures by using decision tree method. The results of experiment demonstrate that the proposed method is pretty efficient to recognize common gestures with a high accuracy.

Motion Planning System for Bin Picking Using 3-D Point Cloud

In this paper, we propose a motion planning system for bin picking using 3-D point cloud. The situation that the objects are put miscellaneously like the inside in a house is assumed. In the home, the equipment which makes an object stand in line doesn’t exist. Therefore the motion planning system which considered a collision problem becomes important. In this paper, Information on the objects is measured by a laser range finder (LRF). The information is used as 3-D point cloud, and the objects are recognized by model-base. We propose search method of a grasping point for two-fingered robotic hand, and propose search method of a path to approach the grasping point without colliding with other objects.

三维激光扫描技术结合3DS MAX在建筑物逆向建模中的应用

针对目前建筑物逆向建模中存在的数据采集效率低、模型精度低、模型效果不真实等问题,提出了一种三维激光扫描技术结合3DS MAX对建筑物进行逆向建模的方法。使用架站式三维激光扫描仪对目标建筑物进行多站扫描,利用SCENE软件对多站点云数据进行配准、去噪生成三维点云,基于三维点云生成建筑物各立面的正射影像,将影像导入AutoCAD并绘制建筑物二维平立面矢量图,将各平立面矢量图导入3DS MAX软件,经建立白模、纹理贴图等过程,生成目标建筑三维模型,最后对模型的精度进行了验证。结果表明,该方法可以高效地建立高精度且具有逼真效果的建筑物三维模型。

3D scene graph prediction from point clouds

Background In this study,we propose a novel 3D scene graph prediction approach for scene understanding from point clouds.Methods It can automatically organize the entities of a scene in a graph,where objects are nodes and their relationships are modeled as edges.More specifically,we employ the DGCNN to capture the features of objects and their relationships in the scene.A Graph Attention Network(GAT)is introduced to exploit latent features obtained from the initial estimation to further refine the object arrangement in the graph structure.A one loss function modified from cross entropy with a variable weight is proposed to solve the multi-category problem in the prediction of object and predicate.Results Experiments reveal that the proposed approach performs favorably against the state-of-the-art methods in terms of predicate classification and relationship prediction and achieves comparable performance on object classification prediction.Conclusions The 3D scene graph prediction approach can form an abstract description of the scene space from point clouds.

A voxel-based fine-scale 3D landscape pattern analysis using laser scanner point clouds

The landscape pattern metrics can quantitatively describe the characteristics of landscape pattern and are widely used in various fields of landscape ecology.Due to the lack of vertical information,2D landscape metrics cannot delineate the vertical characteristics of landscape pattern.Based on the point clouds,a high-resolution voxel model and several voxel-based 3D landscape metrics were constructed in this study and 3D metrics calculation results were compared with that of 2D metrics.The results showed that certain quantifying difference exists between 2D and 3D landscape metrics.For landscapes with different components and spatial configurations,significant difference was disclosed between 2D and 3D landscape metrics.3D metrics can better reflect the real spatial structure characteristics of the landscape than 2D metrics.

Visual perception driven 3D building structure representa tion from airborne laser scanning point cloud

Background Three-dimensional(3D)building models with unambiguous roof plane geometry parameters,roof structure units,and linked topology provide essential data for many applications related to human activities in urban environments.The task of 3D reconstruction from point clouds is still in the development phase,especially the recognition and interpretation of roof topological structures.Methods This study proposes a novel visual perception-based approach to automatically decompose and reconstruct building point clouds into meaningful and simple parametric structures,while the associated mutual relationships between the roof plane geometry and roof structure units are expressed by a hierarchical topology tree.First,a roof plane extraction is performed by a multi-label graph cut energy optimization framework and a roof structure graph(RSG)model is then constructed to describe the roof topological geometry with common adjacency,symmetry,and convexity rules.Moreover,a progressive roof decomposition and refinement are performed,generating a hierarchical representation of the 3D roof structure models.Finally,a visual plane fitted residual or area constraint process is adopted to generate the RSG model with different levels of details.Results Two airborne laser scanning datasets with different point densities and roof styles were tested,and the performance evaluation metrics were obtained by International Society for Photogrammetry and Remote Sensing,achieving a correctness and accuracy of 97.7%and 0.29m,respectively.Conclusions The standardized assessment results demonstrate the effectiveness and robustness of the proposed approach,showing its ability to generate a variety of structural models,even with missing data.