Search Results (36)

Detecting leakage using a point cloud acquired by mobile laser scanning (MLS) presents significant challenges, particularly from within three-dimensional space. These challenges primarily arise from the prevalence of noise in tunnel point clouds and the difficulty in accurately capturing the three-dimensional morphological characteristics of leakage patterns. To address these limitations, this study proposes a classification method based on XGBoost classifier, integrating both intensity and geometric features. The proposed methodology comprises the following steps: First, a RANSAC algorithm is employed to filter out noise from tunnel objects, such as facilities, tracks, and bolt holes, which exhibit intensity values similar to leakage. Next, intensity features are extracted to facilitate the initial separation of leakage regions from the tunnel lining. Subsequently, geometric features derived from the k neighborhood are incorporated to complement the intensity features, enabling more effective segmentation of leakage from the lining structures. The optimal neighborhood scale is determined by selecting the scale that yields the highest F_1-score for leakage across various multiple evaluated scales. Finally, the XGBoost classifier is applied to the binary classification to distinguish leakage from tunnel lining. Experimental results demonstrate that the integration of geometric features significantly enhances leakage detection accuracy, achieving an F_1-score of 91.18% and 97.84% on two evaluated datasets, respectively. The consistent performance across four heterogeneous datasets indicates the robust generalization capability of the proposed methodology. Comparative analysis further shows that XGBoost outperforms other classifiers, such as Random Forest, AdaBoost, LightGBM, and CatBoost, in terms of balance of accuracy and computational efficiency. Moreover, compared to deep learning models, including PointNet, PointNet++, and DGCNN, the proposed method demonstrates superior performance in both detection accuracy and computational efficiency. Full article

High-quality multispectral LiDAR (MSL) data are crucial for land cover (LC) classification. However, the Titan MSL system encounters challenges of inconsistent spatial–spectral information due to its unique scanning and data saving method, restricting subsequent classification accuracy. Existing spectral reconstruction methods often require empirical parameter settings and involve high computational costs, limiting automation and complicating application. To address this problem, we introduce the dual attention spectral optimization reconstruction network (DossaNet), leveraging an attention mechanism and spatial–spectral information. DossaNet can adaptively adjust weight parameters, streamline the multispectral point cloud acquisition process, and integrate it into classification models end-to-end. The experimental results show the following: (1) DossaNet exhibits excellent generalizability, effectively recovering accurate LC spectra and improving classification accuracy. Metrics across the six classification models show some improvements. (2) Compared with the method lacking spectral reconstruction, DossaNet can improve the overall accuracy (OA) and average accuracy (AA) of PointNet++ and RandLA-Net by a maximum of 4.8%, 4.47%, 5.93%, and 2.32%. Compared with the inverse distance weighted (IDW) and k-nearest neighbor (KNN) approach, DossaNet can improve the OA and AA of PointNet++ and DGCNN by a maximum of 1.33%, 2.32%, 0.86%, and 2.08% (IDW) and 1.73%, 3.58%, 0.28%, and 2.93% (KNN). The findings further validate the effectiveness of our proposed method. This method provides a more efficient and simplified approach to enhancing the quality of multispectral point cloud data. Full article

LiDAR point clouds, as direct carriers of 3D spatial information, comprehensively record the geometric features and spatial topological relationships of object surfaces, providing intelligent systems with rich 3D scene representation capability. However, current point cloud semantic segmentation methods primarily extract features through operations such as convolution and pooling, yet fail to adequately consider sparse features that significantly influence the final results of point cloud-based scene perception, resulting in insufficient feature representation capability. To address these problems, a sparse feature dynamic graph convolutional neural network, abbreviated as SFDGNet, is constructed in this paper for LiDAR point clouds of complex scenes. In the context of this paper, sparse features refer to feature representations in which only a small number of activation units or channels exhibit significant responses during the forward pass of the model. First, a sparse feature regularization method was used to motivate the network model to learn the sparsified feature weight matrix. Next, a split edge convolution module, abbreviated as SEConv, was designed to extract the local features of the point cloud from multiple neighborhoods by dividing the input feature channels, and to effectively learn sparse features to avoid feature redundancy. Finally, a multi-neighborhood feature fusion strategy was developed that combines the attention mechanism to fuse the local features of different neighborhoods and obtain global features with fine-grained information. Taking S3DIS and ScanNet v2 datasets, we evaluated the feasibility and effectiveness of SFDGNet by comparing it with six typical semantic segmentation models. Compared with the benchmark model DGCNN, SFDGNet improved overall accuracy $\left(OA\right)$, mean accuracy $\left(mAcc\right)$, mean intersection over union $\left(mIoU\right)$, and $sparsity$ by $1.8\%$, $3.7\%, 3.5\%$, and $85.5\%$ on the S3DIS dataset, respectively. The $mIoU$ on the ScanNet v2 validation set, $mIoU$ on the test set, and $sparsity$ were improved by $3.2\%, 7.0\%$, and $54.5\%$, respectively. Full article

Phenotypic data of cotton can accurately reflect the physiological status of plants and their adaptability to environmental conditions, playing a significant role in the screening of germplasm resources and genetic improvement. Therefore, this study proposes a cotton phenotypic data extraction algorithm that integrates ResDGCNN with an improved region-growing method and constructs a 3D point cloud dataset of cotton covering the entire growth period under real growth conditions. To address the challenge of significant structural variations in cotton organs across different growth stages, we designed an innovative point cloud segmentation algorithm, ResDGCNN, which integrates residual learning with dynamic graph convolution to enhance organ segmentation performance throughout all developmental stages. In addition, to address the challenge of accurately segmenting overlapping regions between different cotton organs, we introduced an optimization strategy that combines point distance mapping with curvature-based normal vectors and developed an improved region-growing algorithm to achieve fine segmentation of multiple cotton organs, including leaves, stems, and flower buds. Experimental data show that, in the task of organ segmentation throughout the entire cotton growth cycle, the ResDGCNN model achieved a segmentation accuracy of 67.55%, with a 4.86% improvement in mIoU compared to the baseline model. In the fine-grained segmentation of overlapping leaves, the model achieved an R² of 0.962 and an RMSE of 2.0. The average relative error in stem length estimation was 0.973, providing a reliable solution for acquiring 3D phenotypic data of cotton. Full article

Effective classification of wood and leaf points from terrestrial laser scanning (TLS) point clouds is critical for analyzing and estimating forest attributes such as diameter at breast height (DBH), above-ground biomass (AGB), and wood volume. To address this, we introduce the Wood–Leaf Classification Network (WLC-Net), a deep learning model derived from PointNet++, designed to differentiate between wood and leaf points within tree point clouds. WLC-Net enhances classification accuracy, completeness, and speed by incorporating linearity as an inherent feature, refining the input–output framework, and optimizing the centroid sampling technique. We trained and evaluated WLC-Net using datasets from three distinct tree species, totaling 102 individual tree point clouds, and compared its performance against five existing methods including PointNet++, DGCNN, Krishna Moorthy’s method, LeWoS, and Sun’s method. WLC-Net achieved superior classification accuracy, with overall accuracy (OA) scores of 0.9778, 0.9712, and 0.9508; the mean Intersection over Union (mIoU) scores of 0.9761, 0.9693, and 0.9141; and F1-scores of 0.8628, 0.7938, and 0.9019, respectively. The model also demonstrated high efficiency, processing an average of 102.74 s per million points. WLC-Net has demonstrated notable advantages in wood–leaf classification, including significantly enhanced classification accuracy, improved processing efficiency, and robust applicability across diverse tree species. These improvements stem from its innovative integration of linearity in the model architecture, refined input–output framework, and optimized centroid sampling technique. In addition, WLC-Net also exhibits strong applicability across various tree point clouds and holds promise for further optimization. Full article

In this paper, an improved point cloud alignment network based on PCRNet is proposed. In the improved model, DGCNN is used as a feature extractor to capture the local and global geometric features of the point cloud, and a fully connected layer is used for feature fusion and rigid-body transformation parameter prediction. Compared with the original PCRNet, the improved network shows higher accuracy and robustness in the point cloud alignment task. In order to verify the performance of the improved network, two classical algorithms, ICP and FGR, are used as benchmarks in our experiment, and experimental comparisons of PCRNet and its improved version are performed under noise-free and noise-containing conditions, respectively. The experimental results show that the improved network (PCRNet+) proposed in our approach outperforms the original PCRNet under different test conditions, including experiments conducted in noise-free, noise-containing, and occlusion scenarios. Specifically, under noise-containing conditions, PCRNet+ surpasses the next-best algorithm, FGR, by over 95.9% across three key metrics. In occlusion scenarios, PCRNet+ achieves more than 100% improvement in all evaluated metrics compared to FGR. Full article

Nowadays, user-generated content has become a crucial channel for obtaining information and authentic feedback. However, due to the varying cultural and educational levels of online users, the content of online reviews often suffers from inconsistencies in specification and the inclusion of arbitrary information. Consequently, the task of extracting key information from online reviews has become a prominent area of research. This paper proposes a combined entity recognition model for online reviews, aiming to improve the accuracy of Named Entity Recognition (NER). Initially, the Non-negative Matrix Factorization (NMF) model is employed to perform thematic clustering on the review texts, and entity types are extracted based on the clustering results. Subsequently, we introduce an entity recognition model utilizing the pre-trained BERT model as an embedding layer, with BiLSTM and DGCNN incorporating residual connection and gating mechanisms as feature extraction layers. The model also leverages multi-head attention for feature fusion, and the final results are decoded using a Conditional Random Field (CRF) layer. The model achieves an F1 score of 86.8383% on a collected dataset of online reviews containing eight entity categories. Experimental results demonstrate that the proposed model outperforms other mainstream NER models, effectively identifying key entities in online reviews. Full article

Semantic segmentation of three-dimensional (3D) plant point clouds at the stem-leaf level is foundational and indispensable for high-throughput tomato phenotyping systems. However, existing semantic segmentation methods often suffer from issues such as low precision and slow inference speed. To address these challenges, we propose an innovative encoding-decoding structure, incorporating voxel sparse convolution (SpConv) and attention-based feature fusion (VSCAFF) to enhance semantic segmentation of the point clouds of high-resolution tomato seedling images. Tomato seedling point clouds from the Pheno4D dataset labeled into semantic classes of ‘leaf’, ‘stem’, and ‘soil’ are applied for the semantic segmentation. In order to reduce the number of parameters so as to further improve the inference speed, the SpConv module is designed to function through the residual concatenation of the skeleton convolution kernel and the regular convolution kernel. The feature fusion module based on the attention mechanism is designed by giving the corresponding attention weights to the voxel diffusion features and the point features in order to avoid the ambiguity of points with different semantics having the same characteristics caused by the diffusion module, in addition to suppressing noise. Finally, to solve model training class bias caused by the uneven distribution of point cloud classes, the composite loss function of Lovász-Softmax and weighted cross-entropy is introduced to supervise the model training and improve its performance. The results show that mIoU of VSCAFF is 86.96%, which outperformed the performance of PointNet, PointNet++, and DGCNN, respectively. IoU of VSCAFF achieves 99.63% in the soil class, 64.47% in the stem class, and 96.72% in the leaf class. The time delay of 35ms in inference speed is better than PointNet++ and DGCNN. The results demonstrate that VSCAFF has high performance and inference speed for semantic segmentation of high-resolution tomato point clouds, and can provide technical support for the high-throughput automatic phenotypic analysis of tomato plants. Full article

Three-dimensional point clouds, as an advanced imaging technique, enable researchers to capture plant traits more precisely and comprehensively. The task of plant segmentation is crucial in plant phenotyping, yet current methods face limitations in computational cost, accuracy, and high-throughput capabilities. Consequently, many researchers have adopted 3D point cloud technology for organ-level segmentation, extending beyond manual and 2D visual measurement methods. However, analyzing plant phenotypic traits using 3D point cloud technology is influenced by various factors such as data acquisition environment, sensors, research subjects, and model selection. Although the existing literature has summarized the application of this technology in plant phenotyping, there has been a lack of in-depth comparison and analysis at the algorithm model level. This paper evaluates the segmentation performance of various deep learning models on point clouds collected or generated under different scenarios. These methods include outdoor real planting scenarios and indoor controlled environments, employing both active and passive acquisition methods. Nine classical point cloud segmentation models were comprehensively evaluated: PointNet, PointNet++, PointMLP, DGCNN, PointCNN, PAConv, CurveNet, Point Transformer (PT), and Stratified Transformer (ST). The results indicate that ST achieved optimal performance across almost all environments and sensors, albeit at a significant computational cost. The transformer architecture for points has demonstrated considerable advantages over traditional feature extractors by accommodating features over longer ranges. Additionally, PAConv constructs weight matrices in a data-driven manner, enabling better adaptation to various scales of plant organs. Finally, a thorough analysis and discussion of the models were conducted from multiple perspectives, including model construction, data collection environments, and platforms. Full article

Traditional deep learning methods employing 2D images can only classify healthy and unhealthy seedlings; consequently, this study proposes a method by which to further classify healthy seedlings into primary seedlings and secondary seedlings and finally to differentiate three classes of seedling through a 3D point cloud for the detection of useful eggplant seedling transplants. Initially, RGB images of three types of substrate-cultivated eggplant seedlings (primary, secondary, and unhealthy) were collected, and healthy and unhealthy seedlings were classified using ResNet50, VGG16, and MobilNetV2. Subsequently, a 3D point cloud was generated for the three seedling types, and a series of filtering processes (fast Euclidean clustering, point cloud filtering, and voxel filtering) were employed to remove noise. Parameters (number of leaves, plant height, and stem diameter) extracted from the point cloud were found to be highly correlated with the manually measured values. The box plot shows that the primary and secondary seedlings were clearly differentiated for the extracted parameters. The point clouds of the three seedling types were ultimately classified directly using the 3D classification models PointNet++, dynamic graph convolutional neural network (DGCNN), and PointConv, in addition to the point cloud complementary operation for plants with missing leaves. The PointConv model demonstrated the best performance, with an average accuracy, precision, and recall of 95.83, 95.83, and 95.88%, respectively, and a model loss of 0.01. This method employs spatial feature information to analyse different seedling categories more effectively than two-dimensional (2D) image classification and three-dimensional (3D) feature extraction methods. However, there is a paucity of studies applying 3D classification methods to predict useful eggplant seedling transplants. Consequently, this method has the potential to identify different eggplant seedling types with high accuracy. Furthermore, it enables the quality inspection of seedlings during agricultural production. Full article

This paper presents a new deep-learning architecture designed to enhance the spatial synchronization between CMOS and event cameras by harnessing their complementary characteristics. While CMOS cameras produce high-quality imagery, they struggle in rapidly changing environments—a limitation that event cameras overcome due to their superior temporal resolution and motion clarity. However, effective integration of these two technologies relies on achieving precise spatial alignment, a challenge unaddressed by current algorithms. Our architecture leverages a dynamic graph convolutional neural network (DGCNN) to process event data directly, improving synchronization accuracy. We found that synchronization precision strongly correlates with the spatial concentration and density of events, with denser distributions yielding better alignment results. Our empirical results demonstrate that areas with denser event clusters enhance calibration accuracy, with calibration errors increasing in more uniformly distributed event scenarios. This research pioneers scene-based synchronization between CMOS and event cameras, paving the way for advancements in mixed-modality visual systems. The implications are significant for applications requiring detailed visual and temporal information, setting new directions for the future of visual perception technologies. Full article

In order to guide orchard management robots to realize some tasks in orchard production such as autonomic navigation and precision spraying, this research proposed a deep-learning network called dynamic fusion segmentation network (DFSNet). The network contains a local feature aggregation (LFA) layer and a dynamic fusion segmentation architecture. The LFA layer uses the positional encoders for initial transforming embedding, and progressively aggregates local patterns via the multi-stage hierarchy. The fusion segmentation module (Fus-Seg) can format point tags by learning a multi-embedding space, and the generated tags can further mine the point cloud features. At the experimental stage, significant segmentation results of the DFSNet were demonstrated on the dataset of orchard fields, achieving an accuracy rate of 89.43% and an mIoU rate of 74.05%. DFSNet outperforms other semantic segmentation networks, such as PointNet, PointNet++, D-PointNet++, DGCNN, and Point-NN, with improved accuracies over them by 11.73%, 3.76%, 2.36%, and 2.74%, respectively, and improved mIoUs over the these networks by 28.19%, 9.89%, 6.33%, 9.89, and 24.69%, respectively, on the all-scale dataset (simple-scale dataset + complex-scale dataset). The proposed DFSNet can capture more information from orchard scene point clouds and provide more accurate point cloud segmentation results, which are beneficial to the management of orchards. Full article

UAV transmission tower inspection is the use of UAV technology for regular inspection and troubleshooting of towers on transmission lines, which helps to improve the safety and reliability of transmission lines and ensures the stability of the power supply. From the traditional manual tower boarding to the current way of manually selecting target camera shooting points from 3D point clouds to plan the inspection path of the UAV, operational efficiency has drastically improved. However, indoor planning work is still labor-consuming and expensive. In this paper, a deep learning-based point cloud transmission tower segmentation (PCTTS) model combined with the corresponding target point localization algorithm is proposed for automatic segmentation of transmission tower point cloud data and automatically localizing the key inspection component as the target point for UAV inspection. First, we utilize octree sampling with unit ball normalization to simplify the data and ensure translation invariance before putting the data into the model. In the feature extraction stage, we encode the point set information and combine Euclidean distance and cosine similarity features to ensure rotational invariance. On this basis, we adopt multi-scale feature extraction, construct a local coordinate system, and introduce the offset-attention mechanism to enhance model performance further. Then, after the feature propagation module, gradual up-sampling is used to obtain the features of each point to complete the point cloud segmentation. Finally, combining the segmentation results with the target point localization algorithm completes the automatic extraction of UAV inspection target points. The method has been applied to six kinds of transmission tower point cloud data of part segmentation results and three kinds of transmission tower point cloud data of instance segmentation results. The experimental results show that the model achieves mIOU of 94.1% on the self-built part segmentation dataset and 86.9% on the self-built instance segmentation dataset, and the segmentation accuracy outperforms that of the methods for point cloud segmentation, such as PointNet++, DGCNN, Point Transformer, and PointMLP. Meanwhile, the experimental results of UAV inspection target point localization also verify the method’s effectiveness in this paper. Full article

Background: Degenerative spinal pathologies are highly prevalent among the elderly population. Timely diagnosis of osteoporotic fractures and other degenerative deformities enables proactive measures to mitigate the risk of severe back pain and disability. Methods: We explore the use of shape auto-encoders for vertebrae, advancing the state of the art through robust automatic segmentation models trained without fracture labels and recent geometric deep learning techniques. Our shape auto-encoders are pre-trained on a large set of vertebrae surface patches. This pre-training step addresses the label scarcity problem faced when learning the shape information of vertebrae for fracture detection from image intensities directly. We further propose a novel shape decoder architecture: the point-based shape decoder. Results: Employing segmentation masks that were generated using the TotalSegmentator, our proposed method achieves an AUC of 0.901 on the VerSe19 testset. This outperforms image-based and surface-based end-to-end trained models. Our results demonstrate that pre-training the models in an unsupervised manner enhances geometric methods like PointNet and DGCNN. Conclusion: Our findings emphasize the advantages of explicitly learning shape features for diagnosing osteoporotic vertebrae fractures. This approach improves the reliability of classification results and reduces the need for annotated labels. Full article

Accurate detection of power fittings is crucial for identifying defects or faults in these components, which is essential for assessing the safety and stability of the power system. However, the accuracy of fittings detection is affected by a complex background, small target sizes, and overlapping fittings in the images. To address these challenges, a fittings detection method based on the dynamic graph convolutional neural network (DGCNN) and U-shaped network (U-Net) is proposed, which combines three-dimensional detection with two-dimensional object detection. Firstly, the bi-level routing attention mechanism is incorporated into the lightweight U-Net network to enhance feature extraction for detecting the fittings boundary. Secondly, pseudo-point cloud data are synthesized by transforming the depth map generated by the Lite-Mono algorithm and its corresponding RGB fittings image. The DGCNN algorithm is then employed to extract obscured fittings features, contributing to the final refinement of the results. This process helps alleviate the issue of occlusions among targets and further enhances the precision of fittings detection. Finally, the proposed method is evaluated using a custom dataset of fittings, and comparative studies are conducted. The experimental results illustrate the promising potential of the proposed approach in enhancing features and extracting information from fittings images. Full article