Cloud Classification Research Articles

Evaluation of classified ground points from National Agriculture Imagery program photogrammetrically derived point clouds

ABSTRACT Studies have shown that digital surface models and point clouds generated by the United States Department of Agriculture’s National Agriculture Imagery Program (NAIP) can measure basic forest parameters such as canopy height. However, all measured forest parameters from these studies are evaluated using the differences between NAIP digital surface models (DSMs) and available lidar digital terrain models (DTMs). A survey of NAIP point cloud classification and related ground point-generated DTMs has not yet been undertaken. This study applies a Support Vector Machine (SVM) to classifying ground and nonground points from NAIP point clouds for test sites in Wyoming and Arizona, USA. Light detection and ranging (lidar) data from the U.S. Geological Survey 3D Elevation Program (3DEP) are used to validate the classified NAIP ground points and their corresponding DTMs. Comparing height differences between filtered NAIP ground points and 3DEP ground points, the SVM classifier’s results show that the vertical root mean square error value is 1.87 m and 1.69 m for the Wyoming and Arizona sites, respectively. If NAIP point clouds were continuously measured, the resulting availability of medium-resolution DTMs would benefit the application of multitemporal forest health monitoring and DTM generation.

The Chalmers Cloud Ice Climatology: retrieval implementation and validation

Abstract. Ice clouds are a crucial component of the Earth's weather system, and their representation remains a principal challenge for current weather and climate models. Several past and future satellite missions were explicitly designed to provide observations offering new insights into cloud processes, but these specialized cloud sensors are limited in their spatial and temporal coverage. Geostationary satellites have been observing clouds for several decades and can ideally complement the sparse measurements from specialized cloud sensors. However, the geostationary observations that are continuously and globally available over the full observation record are restricted to a small number of wavelengths, which limits the information they can provide on clouds. The Chalmers Cloud Ice Climatology (CCIC) is a novel cloud-property dataset that aims to provide an improved climate record of ice hydrometeor concentrations by applying state-of-the-art machine-learning techniques to retrieve ice cloud properties from globally gridded, single-channel geostationary observations that are readily available from 1980 onwards. CCIC offers a novel perspective on the record of geostationary IR observations by providing spatially and temporally continuous retrievals of the vertically integrated and vertically resolved concentrations of frozen hydrometeors, typically referred to as ice water path (IWP) and ice water content (IWC). In addition to that, CCIC provides 2D and 3D cloud masks and a 3D cloud classification. A fully convolutional quantile regression neural network constitutes the core of the CCIC retrieval, providing probabilistic estimates of IWP and IWC. The network is trained against CloudSat retrievals using 3.5 years of global collocations. Assessed on a held-out test dataset, the CCIC-provided IWP and IWC estimates achieve correlations exceeding 0.7 and 0.6, respectively, and biases better than −5 % and −2 % demonstrating considerable skill in estimating both IWP and IWC. In addition, CCIC is extensively validated against both in situ and remote sensing measurements from two flight campaign series and a ground-based radar. The results of this independent validation confirm the ability of CCIC to retrieve IWP and IWC. CCIC thus ideally complements temporally and spatially more limited measurements from dedicated cloud sensors by providing spatially and temporally continuous estimates of ice cloud properties. The CCIC network and its associated software are made accessible to the scientific community.

DMRC-Net: Dual mechanism residual connection network for point cloud analysis

The purpose of point cloud analysis is to extract effective high-dimensional feature representations from point coordinates. However, existing methods only rely solely on global or local features. In this paper, we introduce the Dual Mechanism Residual Connection Network (DMRC-Net), a new end-to-end network architecture for more comprehensive representations. First, local features are encoded using graph convolution and fine-grained attention mechanisms, whereas global features are encoded using a multi-layer perceptron. Second, adaptive residual learning is designed to integrate two encoding results effectively. Furthermore, the structural information in local space is strengthened by collecting low-level geometric correlations. With these designs, DMRC-Net is able to capture both global and local details comprehensively. Experimental results show that our method performs well in point cloud classification, part segmentation, and semantic segmentation tasks.

Classification of 3D Sonar Point Clouds derived Underwater using Machine and Deep Learning (CANUPO and RandLA-Net) Approaches

Classification of 3D Sonar Point Clouds derived Underwater using Machine and Deep Learning (CANUPO and RandLA-Net) Approaches

Advancing Cloud Classification Over the Tibetan Plateau: A New Algorithm Reveals Seasonal and Diurnal Variations

AbstractThe cloud classification algorithm widely used in the International Satellite Cloud Climatology Project (ISCCP) tends to underestimate low clouds over the Tibetan Plateau (TP), often mistaking water clouds for high‐level clouds. To address this issue, we propose a new algorithm based on cloud‐top temperature and optical thickness, which we apply to TP using Advanced Himawari Imager (AHI) geostationary satellite data. Compared with Clouds and the Earth's Radiant Energy System cloud‐type products and ISCCP results obtained from AHI data, this new algorithm markedly improved low‐cloud detection accuracy and better aligned with cloud phase results. Validation with lidar cloud‐type products further confirmed the superiority of this new algorithm. Diurnal cloud variations over the TP show morning dominance shifting to afternoon high clouds and evening mid‐level clouds. Winter is dominated by high clouds, summer by mid‐level clouds, spring by daytime low clouds and nighttime high clouds, and autumn by low and mid‐level clouds.

Neural network-based 3D point cloud detection of targets in unstructured environments

Accurate environmental sensing is an important prerequisite for autonomous driving in off-road environments. Most targets in off-road environments do not have regular shapes, colors, textures and other features, making them difficult to identify. In addition, complex driving conditions can cause large, broadband vibrations in off-road vehicles, which interfere with environment sensing and affect the accuracy and efficiency of perception. To address the above problems, this paper proposes an improved 3D point cloud filtering algorithm for unstructured environments and a point cloud classification method using neural networks, and provides an experimental proof-of-principle of the proposed methods. A comparison of the results under six conditions shows that the amount of data processed by the improved filtering algorithm is 65%–85% of that processed by the conventional filtering algorithm, and the trained neural network model achieves an accuracy of 98.0% and a loss value as low as 0.008 when classifying three typical targets in an unstructured environment. A comparison with algorithms proposed in other papers shows that the proposed method is highly feasible.

Receptive Field Space for Point Cloud Analysis.

Similar to convolutional neural networks for image processing, existing analysis methods for 3D point clouds often require the designation of a local neighborhood to describe the local features of the point cloud. This local neighborhood is typically manually specified, which makes it impossible for the network to dynamically adjust the receptive field's range. If the range is too large, it tends to overlook local details, and if it is too small, it cannot establish global dependencies. To address this issue, we introduce in this paper a new concept: receptive field space (RFS). With a minor computational cost, we extract features from multiple consecutive receptive field ranges to form this new receptive field space. On this basis, we further propose a receptive field space attention mechanism, enabling the network to adaptively select the most effective receptive field range from RFS, thus equipping the network with the ability to adjust granularity adaptively. Our approach achieved state-of-the-art performance in both point cloud classification, with an overall accuracy (OA) of 94.2%, and part segmentation, achieving an mIoU of 86.0%, demonstrating the effectiveness of our method.

PointStaClu: A Deep Point Cloud Clustering Method Based on Stable Cluster Discrimination

Potential inconsistencies between the goals of unsupervised representation learning and clustering within multi-stage deep clustering can diminish the effectiveness of these techniques. However, because the goal of unsupervised representation learning is inherently flexible and can be tailored to clustering, we introduce PointStaClu, a novel single-stage point cloud clustering method. This method employs stable cluster discrimination (StaClu) to tackle the inherent instability present in single-stage deep clustering training. It achieves this by constraining the gradient descent updates for negative instances within the cross-entropy loss function, and by updating the cluster centers using the same loss function. Furthermore, we integrate entropy constraints to regulate the distribution entropy of the dataset, thereby enhancing the cluster allocation. Our framework simplifies the process, employing a single loss function and an encoder for deep point cloud clustering. Extensive experiments on the ModelNet40 and ShapeNet dataset demonstrate that PointStaClu significantly narrows the performance gap between unsupervised point cloud clustering and supervised point cloud classification, presenting a novel approach to point cloud classification tasks.

Fusion of Multiple Basic Element Features for Airborne LiDAR in-house Surveys

When using airborne LiDAR point clouds for city modelling and road extraction, point cloud classification is a crucial step. There are numerous ways for classifying point clouds, but there are still issues like redundant multi-dimensional feature vector data and poor point cloud classification in intricate situations. A point cloud classification method built on the fusing of multikernel feature vectors is suggested as a solution to these issues. The technique employs random forest to classify point cloud data by merging colour information, and it extracts feature vectors based on point primitives and object primitives, respectively. In this study, a densely populated area was chosen as the study area. Light airborne LIDAR mounted on a delta wing was used to collect point cloud data at a low altitude (170 m) over a dense cross-course. The point cloud data were then combined, corrected, and enhanced with texture data, and the houses were vectorized on the point cloud. The accuracy of the results was then assessed. With a median inaccuracy of 4.8 cm and a point cloud data collection rate of 83.3%, using airborne LIDAR to measure house corners can significantly lighten the labour associated with external house corner measurements.This test extracts the texture information of point cloud data through the efficient processing of high-density point cloud data, providing a reference for the application of texture information of airborne LIDAR data and a clear understanding of its accuracy.

Fuzzy preference matroids rough sets for approximate guided representation in transformer

Recently, the transformer has exhibited remarkable performance across various applications, primarily owing to its exceptional capability in capturing global information through the attention mechanism. Nevertheless, the dot-product within attention results in inaccurate and confusing guidance in the presence of uncertain features, subsequently affecting robustness. Hence, introducing an approximate guided method to handle uncertain features is essential for achieving enhanced objectivity. In this study, we employed fuzzy preference relations to construct a matroids structure, introducing it into rough sets to form the fuzzy preference matroids rough sets. Detailed mathematical properties and axiomatic proofs are presented. The closed sets within the matroids form a subspace of the lower approximation proving that the proposed satisfy the approximation ability of rough sets. Further, an approximate guided representation method with a lower approximation of the matroids has been developed. It is integrated into a plug-and-play transformer block that can be flexibly deployed across various tasks. Experimental validation has demonstrated the outstanding performance of this method across 3D point cloud classification, named entity recognition, multimodal emotion recognition, and image classification. Specifically, on the Weibo NER and the IEMOCAP datasets, attain state-of-the-art (SOTA) F1-scores. The resource code is validated at https://github.com/WinnieSunning/FPMRST.

Characterizing clouds with the CCClim dataset, a machine learning cloud class climatology

Characterizing clouds with the CCClim dataset, a machine learning cloud class climatology

Accurately identifying the defects of bubbles and foreign objects under the protective films of electric vehicle batteries by using 3D point clouds

Abstract For the defects of bubbles and foreign objects under the protective film of electric vehicle batteries, it is difficult to accurately identify them over traditional 2D optical images. In this paper, we first propose a supervoxel-based region growing algorithm for pre-segmentation of point clouds. Secondly, we utilize radial basis function interpolation and threshold segmentation methods to accurately segment defect point clouds from the entire point cloud. Finally, we develop a feature descriptor and combine it with support vector machine to classify bubbles and foreign objects under the film. This paper achieves the identification of bubbles and foreign objects under the film through two steps: point cloud segmentation and point cloud classification. Experimental results demonstrate that the proposed point cloud segmentation method exhibits high robustness to noise and the intrinsic curvature of the workpiece. Additionally, in the classification scenario presented in this paper, the proposed feature descriptor outperforms classical feature descriptors. Compared to image-based deep learning methods, the defect recognition algorithm proposed in this paper has clear principles and superior performance, with precision and recall of 95.63% and 96.95%, and an intersection over union metric of 0.926.

Innovative cloud quantification: deep learning classification and finite-sector clustering for ground-based all-sky imaging

Abstract. Accurate cloud quantification is essential in climate change research. In this work, we construct an automated computer vision framework by synergistically incorporating deep neural networks and finite-sector clustering to achieve robust whole-sky image-based cloud classification, adaptive segmentation and recognition under intricate illumination dynamics. A bespoke YOLOv8 (You Only Look Once 8) architecture attains over 95 % categorical precision across four archetypal cloud varieties curated from extensive annual observations (2020) at a Tibetan highland station. Tailor-made segmentation strategies adapted to distinct cloud configurations, allied with illumination-invariant image enhancement algorithms, effectively eliminate solar interference and substantially boost quantitative performance even in illumination-adverse analysis scenarios. Compared with the traditional threshold analysis method, the cloud quantification accuracy calculated within the framework of this paper is significantly improved. Collectively, the methodological innovations provide an advanced solution to markedly escalate cloud quantification precision levels imperative for climate change research while offering a paradigm for cloud analytics transferable to various meteorological stations.

VR map construction for orchard robot teleoperation based on dual-source positioning and sparse point cloud segmentation

The unstructured nature of orchard environments presents significant challenges for autonomous navigation of orchard robots. Teleoperation, combined with virtual reality (VR), has emerged as a promising solution to overcome the limitations of on-board autonomous navigation capabilities in general-purpose robots. However, constructing accurate and semantically meaningful VR maps for orchard environments remains a challenge. This study proposes a novel VR map construction framework for orchard robot teleoperation visualization. It comprises two key components: dual-source combined positioning and semantic segmentation based on sparse point clouds. First, a novel sliding window-based data fusion approach is proposed to address the positioning challenges in semi-obstructed orchard environments. This approach combines GNSS data and laser matching to achieve robust and accurate positioning for orchard robots. Second, the framework capitalizes on the installation characteristics of mobile robot radars to achieve real-time classification of sparse point clouds. This allows for the removal of unstable elements, such as leaves, which would hinder effective teleoperation. By integrating these components, a VR map suitable for VR teleoperation of orchard robots is achieved. To validate the effectiveness of the proposed method, a VR teleoperation platform was constructed. The experimental results demonstrate that the method successfully fulfills the fundamental requirements for map visualization during VR remote operation of orchard robots. Furthermore, this study provides a valuable reference for the application of digital twins in agricultural robotics.

Automatic Classification of Submerged Macrophytes at Lake Constance Using Laser Bathymetry Point Clouds

Submerged aquatic vegetation, also referred to as submerged macrophytes, provides important habitats and serves as a significant ecological indicator for assessing the condition of water bodies and for gaining insights into the impacts of climate change. In this study, we introduce a novel approach for the classification of submerged vegetation captured with bathymetric LiDAR (Light Detection And Ranging) as a basis for monitoring their state and change, and we validated the results against established monitoring techniques. Employing full-waveform airborne laser scanning, which is routinely used for topographic mapping and forestry applications on dry land, we extended its application to the detection of underwater vegetation in Lake Constance. The primary focus of this research lies in the automatic classification of bathymetric 3D LiDAR point clouds using a decision-based approach, distinguishing the three vegetation classes, (i) Low Vegetation, (ii) High Vegetation, and (iii) Vegetation Canopy, based on their height and other properties like local point density. The results reveal detailed 3D representations of submerged vegetation, enabling the identification of vegetation structures and the inference of vegetation types with reference to pre-existing knowledge. While the results within the training areas demonstrate high precision and alignment with the comparison data, the findings in independent test areas exhibit certain deficiencies that are likely addressable through corrective measures in the future.

ModelNet-O: A large-scale synthetic dataset for occlusion-aware point cloud classification

Recently, 3D point cloud classification has made significant progress with the help of many datasets. However, these datasets do not reflect the incomplete nature of real-world point clouds caused by occlusion, which limits the practical application of current methods. To bridge this gap, we propose ModelNet-O, a large-scale synthetic dataset of 123,041 samples that emulates real-world point clouds with self-occlusion caused by scanning from monocular cameras. ModelNet-O is 10 times larger than existing datasets and offers more challenging cases to evaluate the robustness of existing methods. Our observation on ModelNet-O reveals that well-designed sparse structures can preserve structural information of point clouds under occlusion, motivating us to propose a robust point cloud processing method that leverages a critical point sampling (CPS) strategy in a multi-level manner. We term our method PointMLS. Through extensive experiments, we demonstrate that our PointMLS achieves state-of-the-art results on ModelNet-O and competitive results on regular datasets such as ModelNet40 and ScanObjectNN, and we also demonstrate its robustness and effectiveness. Code available: https://github.com/fanglaosi/ModelNet-O_PointMLS.

Global cloud optical depth daily variability based on DSCOVR/ EPIC observations

When investigating the potential effects of cloud changes on climate, the interday and intraday variabilities should be distinguished. Historically, studies have focused on the long-term cloud changes, and the intraday cycles of cloud properties have been mostly ignored partly owing to the limited availability of global datasets to study higher frequency variabilities. In this regard, DSCOVR’s vantage point at the Lagrange L1 point overcomes the temporal limitations of polar orbiters as well as the limited spatial views of geostationary satellites, allowing characterization of the daytime variability of cloud properties using a single sensor. In previous analyses, we used DSCOVR’s EPIC instrument to characterize the diurnal cycles of cloud height and cloud fraction; here, we expand on previous studies to additionally investigate the variability of cloud optical thickness. We observe a recurring diurnal pattern of cloud optical thicknesses for different latitudinal zones that reaches a maximum around noon regardless of the underlying surface. Once we separate the clouds based on their optical thickness into optically thin (0–3), intermediate (3–10), and thick (10–25), we find that these cloud classes follow different optical thickness diurnal cycles on the global scale. We further explore these differences by combining the evolution of cloud fraction and optical thickness of each group to obtain the diurnal evolution of cloud-fraction-weighted optical depths.

Graph Attentive Dual Ensemble learning for Unsupervised Domain Adaptation on point clouds

Due to the annotation difficulty of point clouds, Unsupervised Domain Adaptation (UDA) is a promising direction to address unlabeled point cloud classification and segmentation. Recent works show that adding a self-supervised learning branch for target domain training consistently boosts UDA point cloud tasks. However, most of these works simply resort to geometric deformation, which ignores semantic information and is hard to bridge the domain gap. In this paper, we propose a novel self-learning strategy for UDA on point clouds, termed as Graph Attentive Dual Ensemble learning (GRADE), which delivers semantic information directly. Specifically, with a pre-training process on the source domain, GRADE further builds dual collaborative training branches on the target domain, where each of them constructs a temporal average teacher model and distills its pseudo labels to the other branch. To achieve faithful labels from each teacher model, we improve the popular DGCNN architecture by introducing a dynamic graph attentive module to mine the relation between local neighborhood points. We conduct extensive experiments on several UDA point cloud benchmarks, and the results demonstrate that our GRADE method outperforms the state-of-the-art methods on both classification and segmentation tasks with clear margins.

Fusing differentiable rendering and language–image contrastive learning for superior zero-shot point cloud classification

Zero-shot point cloud classification involves recognizing categories not encountered during training. Current models often exhibit reduced accuracy on unseen categories without 3D pre-training, emphasizing the need for improved precision and interoperability. We propose a novel approach integrating differentiable rendering with contrastive language–image pre-training. Initially, differentiable rendering autonomously learns representative viewpoints from the data, enabling the transformation of point clouds into multi-view images while preserving key visual information. This transformation facilitates optimized viewpoint selection during training, refining the final feature representation. Features are extracted from the multi-view images and integrated into a global multi-view feature using a cross-attention mechanism. On the textual side, a large language model (LLM) is provided with 3D heuristic prompts to generate 3D-specific text reflecting category-specific traits, from which textual features are derived. The LLM’s extensive pre-trained knowledge enables it to capture abstract notions and categorical features relevant to distinct point cloud categories. Visual and textual features are aligned in a unified embedding space, enabling zero-shot classification. Throughout training, the Structural Similarity Index (SSIM) is integrated into the loss function to encourage the model to discern more distinctive viewpoints, reduce redundancy in multi-view imagery, and enhance computational efficiency. Experimental results on the ModelNet10, ModelNet40, and ScanObjectNN datasets demonstrate classification accuracies of 75.68%, 66.42%, and 52.03%, respectively, surpassing prevailing methods in zero-shot point cloud classification accuracy.

DGPCD: a benchmark for typical official-style Dougong in ancient Chinese wooden architecture

Dougong, a distinctive component of ancient Chinese wooden architecture, holds significant importance for the preservation and restoration of such structures. In particular, the northern official-style buildings represent the pinnacle of ancient Chinese construction techniques. In the realm of cultural heritage preservation, the application of deep learning has gradually expanded, demonstrating remarkable effectiveness. Point cloud serving as a crucial source for Dougong, encapsulates various information, enabling support for tasks like Dougong point cloud classification and completion. The quality of Dougong datasets directly impacts the outcomes of DNNs (deep neural networks), as they serve as the foundational data support for these models. The typical official-style Dougong, with its standardized and repetitive structural patterns, is highly suitable for training DNNs to accurately recognize and analyze these complex architectural elements. However, due to the inherent characteristics of Dougong, such as coplanarity and occlusion, acquiring point cloud data is challenging, resulting in poor data quality and organizational difficulties. To address this issue, our study adopts a multi-source data fusion approach to tackle the challenges of insufficient data quantity and poor data quality. Further, through data augmentation, we enhance the dataset’s volume and generalize its characteristics. This effort culminates in the creation of the typical official-style Dougong Point Cloud Dataset (DG Dataset), poised to support deep learning tasks related to Dougong scenarios.