搜索结果：lidar

Conventional ex vivo molecular assessment of nodal specimens lacks three-dimensional (3D) spatial context, making localisation of tracer-avid sentinel nodes (SNs) within large en bloc inguinal lymph node dissection (ILND) specimens challenging. We evaluated the technical feasibility of combining light detection and ranging (LiDAR)-registered freehand SPECT (fhSPECT) with near-infrared (NIR) fluorescence imaging for spatially augmented localisation of hybrid tracer uptake in solitary SNs and SNs residing within ILND specimens. Fourteen patients with penile cancer (25 cN0 groins; 3 cN1 groins) underwent radio- and fluorescence-guided lymph node (LN) surgery following preoperative lymphoscintigraphy and SPECT/CT. Resected specimens underwent conventional ex vivo assessment, followed by LiDAR-registered fhSPECT with co-registered NIR-fluorescence imaging. These 3D models were compared with conventional ex vivo assessments, preoperative imaging, and histopathological outcomes. In total, 40 solitary tracer-avid LNs and 3 ILND specimens underwent ex vivo assessment. LiDAR/fhSPECT localised radioactive hotspots in 38/40 solitary LNs, consistent with conventional gamma probe assessment, while both conventional and LiDAR-registered NIR-fluorescence imaging detected fluorescent signal in all 40 LNs. Among the 3 ILND procedures, preoperative SPECT/CT identified 6 tracer-avid hotspots within the anatomical resection template. Conventional gamma probe and NIR-fluorescence detected only 4 and 5 hotspots, respectively, whereas LiDAR/fhSPECT localised all 6 hotspots in the corresponding resected specimens. Histopathology identified 6 tumour-positive LNs, all of which were hybrid tracer-avid and effectively localised by LiDAR/fhSPECT. LiDAR-registered 3D surface mapping provided anatomical context to radioactive and fluorescent molecular imaging signals, enabling localisation of tracer-avid LNs in complex ex vivo ILND specimens. These findings support the feasibility of spatially augmented localisation of hybrid tracer uptake and warrant validation in larger studies.

Point cloud images from LiDAR often suffer distortion due to platform vibration. This paper proposes a LiDAR-INS (Inertial Navigation System) integrated navigation method to address the challenge of low positioning accuracy in complex environments. To solve problems like GPS signal denial and vibration interference, we present a method for achieving centimeter-level positioning. This method uses INS attitude information to compensate for LiDAR vibration errors. A vibration error model is established to quantify the impact of vibration on point cloud distortion. High-frequency INS attitude data is then used to correct the LiDAR point cloud distortion caused by platform vibration. Leveraging the non-repetitive scanning pattern of prism-based LiDAR, a joint compensation strategy for vibration error and angular error is proposed. This strategy enhances both point cloud density and positioning robustness. State and observation equations for the LiDAR-INS integrated navigation system are derived. A Kalman filter is employed to achieve optimal data fusion between the LiDAR and INS. Finally, field experiments were conducted in both laboratory settings and a typical application scenario: a tunnel construction site. These experiments validate the effectiveness of the proposed method.

Autonomous driving systems face critical perception failures in dense fog, where conventional RGB cameras suffer from severe degradation due to atmospheric scattering and reduced visibility. This paper presents an adaptive multi-modal fusion framework that synergistically integrates gated imaging with 3D LiDAR point clouds to achieve robust obstacle detection under visibility conditions as low as 50 m. Unlike standard cameras that passively capture scattered ambient light, gated cameras employ time-synchronized active illumination to physically filter backscattered photons, preserving structural features even in low-visibility scenarios. We propose a novel Adaptive Feature-Weighting Network (AFW-Net) that dynamically adjusts sensor modality contributions based on real-time environmental degradation assessment. The framework incorporates three key innovations: (1) a cross-modal feature extraction module that exploits the complementary physical properties of gated imaging and LiDAR, (2) an attention-based adaptive fusion mechanism that quantifies per-modality reliability through uncertainty estimation, and (3) a degradation-aware training strategy using weather-specific augmentation. Extensive experiments on the Princeton Automated Driving Dataset demonstrate that our approach maintains detection average precision (AP) above 82% under dense fog conditions (50 m visibility), representing a 23.7% improvement over state-of-the-art RGB-LiDAR fusion methods that exhibit substantial performance degradation to 58.4% AP. Ablation studies validate the necessity of each component, and cross-dataset evaluation confirms the generalization capability of the proposed framework. The adaptive weighting mechanism proves particularly effective, dynamically rebalancing modality contributions across the gated imaging and LiDAR branches while maintaining LiDAR geometric constraints. This work establishes a robust perception paradigm for safety-critical autonomous systems operating in low-visibility environmental conditions.

Due to the differences in the operational principles and target objects, the coherent Doppler wind Lidar (CDWL) and direct detection Doppler wind Lidar (DDWL) each have their own advantages and disadvantages. To fully leverage the advantages of the two types of wind lidar, a hybrid Doppler wind Lidar (HDWL) named WindMast Trop 15 K, combining coherent-detection and direct-detection schemes, is developed. A data fusion method designed for the HDWL based on real-time fitting prediction (RTFP) is proposed herein. We conducted long-term synchronous observation experiments at the Beijing Nanjiao Meteorological Observatory from June to September 2025 to evaluate the wind measurement accuracy of the HDWL. And three measurement cases were presented to demonstrate the data fusion method's performance. The results indicate that: 1) the HDWL coherent-detection module measurement data are consistent well with the radiosonde measurements, with the correlation coefficient (R2) of wind speed (wind direction) reaching 0.969 (0.993), and the root mean square error (RMSE) is 0.632 m·s-1 (8.209°). Since the aerosol load above 3 km becomes lower, the data acquisition rate of the coherent-detection module measurements decreases accordingly, and the wind speed error gets larger; 2) The consistency between the HDWL direct-detection module and radiosonde varies within different altitude ranges. Below 2 km, the high concentration and rapid variation of aerosol distribution cause a significant wind speed deviation. Subsequently, with the altitude increasing, the RMSE of wind speed decreases to a minimum of 0.98 m·s-1 and then increases, while the RMSE of wind direction decreases and stabilizes around 5°; and 3) Compared with the pre-fusion data, the measurement accuracy of wind speed and direction is improved to varying degrees after fusion. Taking the case on September 20, 2025,local time, as an example, the RMSE of wind speed decreased from 0.891 m·s-1 to 0.506 m·s-1, with a reduction of 43.2%. The RMSE of wind direction decreased from 4.663° to 3.889°, with a drop of 16.6%. The results verify the effectiveness of the RTFP method.

LiDAR-based perception in autonomous systems is fundamentally limited by sparse vertical sampling and further degraded by structured beam dropout caused by occlusions, sensor faults, or reduced-cost LiDAR hardware. These degradations disrupt vertical geometric continuity and negatively affect downstream perception tasks such as object detection, localization, and scene understanding. Existing reconstruction approaches often struggle to balance reconstruction accuracy with the computational efficiency required for real-time autonomous operation. This paper presents SuperiorGAT, a graph attention-based framework for reconstructing missing elevation information in sparse LiDAR point clouds under structured beam loss. The proposed method models LiDAR scans as beam-aware graphs and enhances standard graph attention networks using gated residual fusion and lightweight feed-forward refinement to improve vertical reconstruction fidelity without increasing network depth. The effectiveness of SuperiorGAT is evaluated on multiple KITTI environments, including Person, Road, Campus, and City, as well as through cross-dataset validation on nuScenes with lower vertical resolution. Additional experiments under severe structured sparsity further evaluate robustness in a 16-beam-equivalent sensing condition. Results demonstrate that SuperiorGAT achieves lower overall reconstruction error and improved geometric consistency compared to interpolation-based methods, PointNet-based models, and standard GAT baselines while maintaining computational efficiency suitable for real-time perception pipelines.

Recent advances in roadside sensing technologies, including camera-based systems, radar, and LiDAR, have enabled high-resolution sampling of vehicle trajectories, overcoming the temporal and spatial limitations of traditional data collection methods. Among these, LiDAR sensing has been widely adopted for traffic monitoring and surrogate safety analysis due to its high spatial accuracy and temporal resolution. However, sensor noise and occlusion in roadside LiDAR frequently introduce tracking point offsets and trajectory discontinuities, reducing the reliability of vehicle counts, traffic state estimation, and conflict analysis. To address these challenges, this study proposes a post-processing method based on time-space analysis to detect and correct occlusion-induced trajectory discontinuities. By exploiting the inherent spatiotemporal consistency of vehicle movements, the proposed approach identifies fragmented trajectories, reconstructs continuous vehicle paths, and recovers realistic traffic patterns. Validated on real-world LiDAR data collected at an urban intersection in Reno, Nevada, across four 30 min traffic periods covering AM and PM peak conditions on weekdays and weekends, the proposed method achieves an average precision of 0.989 and an average F1-score of 0.948, outperforming IMM, GNN-RM, and HMM + Viterbi benchmark methods. Count accuracy improved from 85.5% to 97.4% across all evaluated periods, confirming the method's effectiveness under occlusion conditions.

Roadside LiDAR is a key sensing technology for intelligent transportation systems (ITSs) due to its high-precision spatial information and reliable monitoring of traffic environments. However, extracting traffic information from LiDAR point cloud data remains challenging because measurements are produced through angular sampling, causing the spacing between adjacent points to depend on radius and beam distribution. This study proposes a geometry-aware framework that incorporates LiDAR sampling geometry into the neighborhood criterion used to determine point-to-point association. The formulation defines neighborhood tolerance as a function of radial distance and vertical angular separation, enabling clustering decisions that are consistent with the sensing mechanism. In addition, the approach integrates deployment constraints based on sensor mounting height and region-of-interest limits to maintain physically meaningful connectivity under roadside sensing conditions. A systematic calibration procedure is conducted to estimate the scaling factor and radial spacing parameters and evaluate the method using both controlled and real-world datasets. Experimental results reveal that the proposed approach improves clustering accuracy and stability by reducing false negatives in sparse regions while avoiding excessive cluster merging in dense areas. The method demonstrates robust performance across varying sensing conditions and achieves higher accuracy than baseline approaches without parameter retuning, while introducing negligible computational overhead.

Accurate forest aboveground carbon (AGC) storage estimation is critical for the global carbon cycle and ecosystem sequestration. Optical data suffer from spectral saturation, and spaceborne LiDAR from spatial discontinuity, both hindering seamless high-precision mapping. This study integrated TECIS LiDAR data with Sentinel-2 imagery to construct three-dimensional enhanced spectral indices (TDESIs) by pixel-wise fusing forest canopy height with optical spectral, textural, and fractional vegetation cover features for machine-learning-based AGC estimation. TDESIs outperformed optical-only models, increasing R2 from 0.666 to 0.750 and reducing RMSE by approximately 14.8%, while alleviating spectral saturation. The best performance was achieved by a random forest model using TDESIs and LiDAR-derived metrics (R2 = 0.764, RMSE = 10.728 Mg C/ha). The resultant AGC map (range 8.03-95.69 Mg C/ha) closely matched field conditions, with enhanced discriminability in high-AGC forest. The TDESIs method enables efficient multi-source data integration for large-scale, high-accuracy AGC monitoring.

Reliable light detection and ranging (LiDAR) is often challenged by multiscale water droplets on protective covers, as macroscopic raindrops cause refraction and diffraction, while microscopic fog condensation scatters light, distorting returned signals. Here, we present a bioinspired strategy for multiscale droplet clearance using plasmonic nanocomposite helices that couple passive photothermal antifogging with pressure-stable hydrophobic water repellence. Inspired by penguin feather barbules, copper nanoparticles are embedded within three-dimensional silica nanohelices via glancing angle co-deposition. The copper nanoparticles provide visible-light plasmonic heating while preserving >85% transmittance at 905 nm, and the helical architecture imparts hierarchical roughness, yielding a 143° contact angle. Under 1 sun illumination, the surface shows a 9.3°C temperature rise, clearing condensation within 6 seconds. During natural rainfall (3-5 mm/h), LiDAR transmission through the plasmonic nanohelices remains at 100%, whereas bare glass drops to 70% after 5 minutes. These results demonstrate weather-adaptive LiDAR operation for robust autonomous sensing.

Bird's-eye-view (BEV) LiDAR detectors provide an efficient representation for real-time 3D perception, but their confidence scores may be imperfectly aligned with the localization quality of decoded 3D boxes. This score-localization mismatch can reduce the reliability of detection ranking, especially under strict IoU criteria and range-dependent LiDAR observations. To address this issue, we propose Spatially-Aware Quality Calibration (SAQC), a lightweight reliability-oriented scoring framework for BEV LiDAR 3D vehicle detection. SAQC estimates localization quality from coordinate-augmented local BEV feature patches around detected object centers, fuses the estimated quality with the raw detector score for quality-aware ranking, and applies post hoc sigmoid calibration with soft IoU targets for numerical quality alignment. Experiments on the KITTI Car validation split using an SFA3D-style center-based detector show that SAQC improves Moderate 3D AP from 89.13 to 89.57 at IoU = 0.7 and from 74.40 to 75.72 at IoU = 0.8. It also increases score-IoU Spearman correlation from 0.3685 to 0.4043 and reduces post-calibration Q-ECE from 0.0284 to 0.0183, while maintaining 103.6 FPS. These results indicate that local BEV spatial context can improve score-localization reliability without modifying decoded box geometry.

Accurate vehicle localization must be maintained even in tunnel sections where GNSS reliability is degraded. However, conventional GNSS/INS-based localization rapidly accumulates errors in such environments, affecting lane-level decision-making and path-following stability. To address this problem, this study proposes a dedicated localization support sign for stable LiDAR observation and a point-cloud-registration-based correction algorithm. The proposed method detects a dedicated sign using a PointPillars-based detector, and the corresponding point cloud is registered to a pre-built reference map to estimate a rigid correction transform online. The sign was installed in a tunnel section of a proving ground that reproduces real-road conditions. For evaluation, the driving sequence was analyzed by separating the pre-entry section, the tunnel section before dedicated-sign recognition, and the section after dedicated-sign recognition. The proposed pipeline substantially reduced localization error after dedicated-sign recognition, compared with the GNSS/INS-only baseline. The dedicated sign also provided more stable correction than ordinary tunnel structures within the same registration pipeline. These results indicate that the proposed LiDAR-based pipeline can suppress localization drift in GNSS-degraded sections.

Plant height during the early growth stage of rice is a key indicator reflecting canopy establishment rate, tillering potential, and overall growth vigor, all of which critically determine final yield formation. Conventional manual measurements fail to capture high-frequency, continuous, and non-destructive monitoring of plant height dynamics, limiting the understanding of early growth vigor and its genetic mechanisms. In this study, a UAV-based LiDAR system was employed to acquire canopy point clouds of 211 rice accessions across ten time points within 40 days after transplanting. High-resolution canopy height models (CHMs) were generated, and continuous plant height trajectories H(t) were reconstructed using piecewise cubic Hermite interpolation (PCHIP). The first derivative V(t) quantified growth rate dynamics and identified the timing of maximum growth (Tmax), enabling precise differentiation of early growth patterns among geno-types. Genome-wide association analysis (GWAS) using a mixed linear model (MLM, Q+K) detected 604 significant SNPs, among which 33 were stably expressed across environments. Five candidate genes were identified within ±200 kb windows, mainly encoding proteins related to cell elongation, hormone signaling, and photosynthetic metabolism. The results highlight that LiDAR-based dynamic monitoring of plant height, coupled with genomic association analysis, provides a robust framework for quantifying rice early growth vigor and elucidating its molecular basis, offering valuable guidance for breeding high-vigor "early-establishing" rice cultivars.

Accurate tree counting from remote sensing data is essential for forest inventory, biomass estimation, carbon accounting, and ecological monitoring. However, existing approaches predominantly rely on airborne RGB imagery and often struggle in complex forest scenes where neighboring crowns exhibit highly similar textures and colors and where overlapping crown boundaries become ambiguous. To address this limitation, the LiDAR-derived Canopy Height Model (CHM) is introduced as a complementary modality that provides explicit cues on canopy height variation and vertical structure to support RGB-based analysis. Building on this, we propose BCAR-Net, a broker-guided RGB and depth (RGB-D) multimodal framework that couples bidirectional cross-modal interaction, adaptive tri-branch fusion, and auxiliary reconstruction within a two-stage optimization scheme. Specifically, a bidirectional cross-attention U-Net generates an intermediate broker RGB-D representation from paired RGB images and depth maps through symmetric bidirectional cross-attention between the two modalities and direction-aware gating. The original RGB image, depth map, and broker representation are then jointly encoded by three weight-sharing branches and adaptively aggregated by a spatial fusion gate for density-map regression. To regularize the fused latent feature, a multi-scale cross-attention reconstruction decoder provides auxiliary RGB and depth reconstruction supervision by querying multi-scale BCA-UNet encoder features through 2D cross-attention, and a reconstruction-oriented first stage replaces externally generated fused-image supervision, yielding a task-consistent optimization scheme. Experiments on the NEONTreeEvaluation benchmark show that BCAR-Net consistently outperforms single-modality settings and direct RGB-D concatenation multimodal baseline. Additional experiments on a public UAV RGB-LiDAR dataset provide a small-scale supplementary evaluation under a different acquisition setting, where BCAR-Net achieves modest but consistent improvements over RGB-only and depth-only baselines. These results demonstrate that the proposed framework offers an effective but computationally cautious solution for tree counting in complex forest environments.

Accurate and wide-area estimation of seaweed biomass is essential for evaluating blue carbon. Conventional diver surveys and two-dimensional (2D) aerial imagery analysis face challenges such as intensive labor and biomass underestimation. While Unmanned Aerial Vehicle-based Light Detection and Ranging (UAV-LiDAR) provides dense 3D spatial data, classifying point clouds in extremely shallow coastal waters with dense kelp and artificial structures remains difficult. This study establishes a high-accuracy biomass estimation method using UAV-LiDAR and PointNet. A heuristic hybrid filtering approach combining physical constraints and local statistics was developed to automatically generate high-quality reference data. The trained PointNet successfully segmented complex point clouds into four classes with an overall accuracy of 94.2%. To calculate biomass, we introduced a volume correction model based on point cloud density (coverage) to mitigate overestimation caused by internal canopy gaps. This correction yielded estimated wet weights nearly identical to the in situ measurements (an approximate 3% difference), confirming highly accurate biomass reproduction. Furthermore, while the conventional 2D maximum likelihood method underestimated total biomass, our 3D point cloud analysis successfully quantified the dense, overlapping canopy. This framework significantly improves the efficiency and accuracy of blue carbon monitoring.

This study conducted a six-month time-series micro-topographic analysis using high-resolution UAV LiDAR technology to precisely characterize the complex terrain changes in regressive tidal creeks within coastal wetlands. To overcome the unique challenges posed by vegetation-dense regressive tidal flats, the LiDAR Penta Return (5-pulse) mode was applied, yielding high-density point cloud data with an average of 174 pts/m2. The analysis successfully reproduced the bare earth surface beneath the vegetation canopy at sub-centimeter-level precision, overcoming the limitations of conventional optical surveying, and enabled quantitative detection of micro-topographic changes of ±25 cm or greater. Time-series analysis based on the DEM of Difference (DoD) revealed spatiotemporally asymmetric erosion and deposition patterns concentrated at the lower elevation zone (0.0-2.0 m) and slope boundaries of the regressive tidal creek. However, the apparent large elevation changes in the lowest, intermittently inundated creek-bed zone (including a maximum of about 3.7 m between the summer surveys, T2-T1) were found to scale monotonically with the tide level at the time of each flight, indicating that they are governed by the tide-dependent water-surface return rather than by genuine bed erosion. After excluding this water-affected zone, the consistently sub-aerial surface showed only modest net change over the six-month period, indicating that the regressive tidal creek adjusts gradually rather than through abrupt large-magnitude erosion and deposition. This study presents the essential value of high-precision time-series monitoring for assessing the geomorphic stability of coastal wetlands in an environment where extreme weather events under climate change are increasing in frequency.

Accurate localization is a critical requirement for autonomous vehicle (AV) navigation, particularly in environments where GPS signals are unreliable or unavailable. A wide range of LiDAR-based point cloud similarity metrics have been proposed for high-definition (HD) map localization, but systematic comparisons of distinct metric families on the same real-world dataset, under identical conditions, remain scarce. In this paper, we present an offline comparative study of three point cloud similarity metrics within a unified HD map-based localization framework, under the assumption of largely static environments and planar, yaw-dominant vehicle motion typical of on-road driving. The HD map is constructed as a directed graph of GPS coordinates, each linked to a corresponding LiDAR scan, collected over a 30-min drive on a university campus using a Velodyne VLP-16 sensor and a ublox ZED-F9P RTK-GPS receiver, yielding 19,500 time-synchronized point clouds. Within this framework, we develop and compare three similarity metrics drawn from distinct families: Fast Point Feature Histograms (FPFH) with KDTree-based matching, Procrustes-based alignment via singular value decomposition, and a planar projection method based on 2D angular histogram cross-correlation. Each metric is evaluated on the same dataset in terms of similarity score profile (localizability) and per-pair computational cost. FPFH provides rich local geometric matching but at an average per-pair cost of approximately 1018 s, making it suitable only for offline analysis. Procrustes alignment yields the smoothest score profiles, with an exact self-similarity baseline of zero, at an average of 2.63 s per pair. The planar projection method produces the most location-invariant profiles at an average of 11.6 s per pair. We also discuss the recursive localization architecture into which any of these metrics could be embedded, and analyze the gap between current per-pair costs and what would be required for online deployment, which we identify as a direction for future work. The study contributes a controlled, reproducible benchmark of three metric families on a single real-world dataset, and provides guidance for selecting similarity metrics under stated operating assumptions.

This study evaluates an Adaptive Monte Carlo Localization-Extended Kalman Filter (AMCL-EKF) pose-estimation stack for repeatable 2D LiDAR SLAM in GPS-denied indoor inspection scenarios. AMCL was used as an online map-referenced correction source fused with LiDAR odometry and Inertial Measurement Unit (IMU) data, and the resulting pose estimate was supplied online to three SLAM backends: Cartographer, GMapping, and SLAM Toolbox. Experiments were performed with a wheeled Husarion Panther and a quadruped Boston Dynamics Spot in three indoor environments of different geometric complexity, producing 720 SLAM executions. Trajectory repeatability was assessed using SE(2)-aligned pairwise and centroid-based ATE-style dispersion and translational RPE, while map repeatability was evaluated with occupied-cell IoU. Accordingly, the metrics were used to quantify between-run dispersion rather than absolute accuracy against external ground-truth data. The results show that AMCL-EKF fusion is highly dependent on the environment, platform, and SLAM backend. AMCL improved selected configurations, especially for Spot in structured environments and for Panther map consistency, but degraded others in geometrically repetitive corridors and mixed-structure spaces. The study also shows that the presence of AMCL-assisted odometry correction alone does not determine final trajectory repeatability, because each SLAM backend incorporates the supplied fused pose estimate differently. The findings support confidence-aware AMCL integration and motivate integrated SLAM architectures resistant to over-correction. These results provide guidance for robust autonomous mapping and inspection with heterogeneous mobile robotic platforms in real environments.

Deep learning-based fusion of hyperspectral images (HSI) and LiDAR has achieved strong performance in multimodal remote sensing classification, but its success is heavily constrained by the high cost of pixel-wise annotation. In extremely label-scarce regimes, such as 2-5 labeled samples per class, conventional deep models are prone to severe overfitting, while standard semi-supervised learning (SSL) methods often suffer from confirmation bias because pseudo-labels are generated from unstable early-stage representations. To address these challenges, we propose Prototype-Guided Progressive Learning (PGPL), a unified framework for few-shot HSI-LiDAR classification. Instead of relying solely on model confidence in latent space, PGPL first constructs a reliable initialization pool directly in the original data domain using spectral-angle and elevation-consistency cues, and then progressively expands the training set through class-balanced pseudo-label admission and temporal confidence stabilization. In this way, the framework improves pseudo-label reliability during both initialization and subsequent self-training. Extensive experiments on three benchmark datasets demonstrate that PGPL consistently outperforms state-of-the-art supervised and semi-supervised baselines under the corresponding 2-5-shot settings, achieving overall accuracy gains of 4.64% points on Houston, 1.16% on Trento, and 3.92% on MUUFL over the strongest competing methods, while also yielding higher pseudo-label purity. The source code will be publicly available at https://github.com/zhangyiyan001/PGPL.

High-quality 3D perception is essential for autonomous vehicles, urban analytics, and the development of intelligent transportation systems. However, existing LiDAR datasets are limited in their representation of fine-grained roadway and pedestrian infrastructure, and geographic diversity, particularly for environments common in North American cities. This paper introduces YEG3D, a large-scale, point-wise annotated mobile laser scanning (MLS) dataset comprising more than 682 million points collected across 14 km of urban roadway in Edmonton, Canada. The dataset includes a fine-grained taxonomy of 18 semantic classes, with an emphasis on detailed pedestrian, cyclist, and roadway infrastructure rarely distinguished in existing benchmarks. We additionally present a comprehensive baseline evaluation using five state-of-the-art semantic segmentation models, including PointNet++, DGCNN, KPConv, KPConvX, and Point Transformer v3. Among the evaluated models, Point Transformer V3 achieves the strongest overall performance, attaining 81.8% overall accuracy, 46.2% mean Intersection over Union (mIoU), and 56.8% mean F1 score, outperforming all other architectures across both global and class-level metrics. Detailed confusion matrix analysis reveals that while large structural classes are segmented reliably, fine-grained elements such as markings, bike lanes, and crosswalks remain challenging due to sparsity, occlusion, and class imbalance. YEG3D provides a new foundation for advancing research in 3D semantic segmentation, urban perception, and infrastructure-aware autonomous systems, and will be expanded in future releases to broaden its geographic and semantic coverage.