<rss version="2.0" xmlns:atom="https://www.w3.org/2005/Atom">
  <channel>
    <title>Transport Research International Documentation (TRID)</title>
    <link>https://trid.trb.org/</link>
    <atom:link href="https://trid.trb.org/Record/RSS?s=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" rel="self" type="application/rss+xml" />
    <description></description>
    <language>en-us</language>
    <copyright>Copyright © 2026. National Academy of Sciences. All rights reserved.</copyright>
    <docs>http://blogs.law.harvard.edu/tech/rss</docs>
    <managingEditor>tris-trb@nas.edu (Bill McLeod)</managingEditor>
    <webMaster>tris-trb@nas.edu (Bill McLeod)</webMaster>
    <image>
      <title>Transport Research International Documentation (TRID)</title>
      <url>https://trid.trb.org/Images/PageHeader-wTitle.jpg</url>
      <link>https://trid.trb.org/</link>
    </image>
    <item>
      <title>DualSeg26: Instance segmentation of ice and snow thickness sections from shipborne imagery with task-decoupled feature pyramids</title>
      <link>https://trid.trb.org/View/2785488</link>
      <description><![CDATA[Shipborne near-field optical imagery directly captures visible ice and snow thickness sections during polar navigation, but these targets are often thin, elongated, low-contrast, and locally ambiguous. Adjacent ice pieces may expose touching sections of the same class, making semantic segmentation prone to region merging when instance-level separation is required. We therefore formulate visible thickness-section delineation as an instance segmentation problem and propose DualSeg26, a YOLO26-based task-decoupled framework whose core design combines separate detection and segmentation necks with a conservative near-identity cross-stream coupling module adapted from manifold-constrained hyper-connections. A semantic P2 head-only branch provides supporting high-resolution refinement for mask prediction. Experiments on a 511-image shipborne dataset show that the final model achieves 0.498 Box AP50, 0.251 Box AP50:95, and 0.267 Mask AP50:95, improving Mask AP50:95 by 0.043 over YOLO26x-Seg. A three-run reproducibility check also gives the highest average AP among the main ablation variants. The resulting segmentation masks can support downstream functions, such as estimating physical ice thickness from calibrated image geometry.]]></description>
      <pubDate>Wed, 07 Oct 2026 16:48:48 GMT</pubDate>
      <guid>https://trid.trb.org/View/2785488</guid>
    </item>
    <item>
      <title>Beaufort-geometry ordinal learning for cross-domain sea-state image classification</title>
      <link>https://trid.trb.org/View/2782448</link>
      <description><![CDATA[Visual sea-state perception is a fundamental capability for intelligent maritime systems. However, most deep image classifiers treat Beaufort-scale recognition as a nominal multi-class problem. This formulation overlooks the ordered structure of the Beaufort scale and the distinction between adjacent errors and distant displacements. This study proposes a Beaufort-geometry ordinal cumulative distribution framework for cross-domain sea-state image classification. The method embeds Beaufort states into an ordered geometric label space and optimizes cumulative probability consistency across Beaufort boundaries. It matches predicted and target cumulative distributions, penalizing probability mass transported across multiple sea-state levels. The framework is evaluated using eight domain-separated protocols covering in-domain recognition, shipborne cross-domain transfer and external-domain transfer to MU-SSiD. The comparison includes six public-backbone baselines: ResNet-101, MobileNetV2, ViT-B/32, DeiT-III, Swin-T and XCiT-S. Four additional Beaufort-aware ordinal objectives are included for family-level ablation. Across the eight protocols, the proposed method achieves the highest mean ordinal consistency score of 0.898, exceeding all evaluated baselines and ordinal-loss variants. Sample-level and error-distance analyses show that it converts selected severe Beaufort displacements into exact or adjacent predictions. It also improves local ordinal consistency under shipborne cross-domain transfer and suppresses selected high-state false predictions in external low-state cases. These results show that incorporating Beaufort geometry into cumulative distribution learning supports ordinally consistent visual sea-state recognition under domain shift.]]></description>
      <pubDate>Thu, 01 Oct 2026 11:44:53 GMT</pubDate>
      <guid>https://trid.trb.org/View/2782448</guid>
    </item>
    <item>
      <title>Research on High-Resolution Imaging System with the Big Field of View Based on the Artificial Compound Eye for Overhead Contact Line System Intelligent Detection</title>
      <link>https://trid.trb.org/View/2580390</link>
      <description><![CDATA[For high-speed railway Overhead Contact line System (OCS) in operation, maintenance and management of image acquisition and background information processing problems, such as: small image acquisition detector range, insufficient image resolution and low integration, etc. Research on intelligent recognition algorithms of OCS typical characteristics based on image splicing is to improve the level of intelligent and informatized operation, maintenance management. Use a deep learning detection method for identifying small targets in a complex background shared by multi-scale feature weights, and a new generation of intelligent computing platform, to break through the existing speed of image feature information processing, innovative improved image stitching algorithm, and improve the stitching quality of artificial compound eye images, for building a set of high resolution imaging system with the big field of view based on the artificial compound eye to further improve high-speed railway Overhead Contact line system(OCS) detection. The research results provide systematic ideas and important guarantees for the operation and maintenance management design of high-speed railway Overhead Contact line system (OCS).]]></description>
      <pubDate>Wed, 30 Sep 2026 16:24:50 GMT</pubDate>
      <guid>https://trid.trb.org/View/2580390</guid>
    </item>
    <item>
      <title>Mutual Distillation Attribute Fusion Network for Multimodal Vehicle Object Detection</title>
      <link>https://trid.trb.org/View/2780896</link>
      <description><![CDATA[Multimodal vehicle object detection aims to enhance autonomous driving system's adaptability to complex real-world scenarios by integrating multimodal images. However, the presence of multimodal noise interference and the challenge of attribute entanglement hinder the performance improvement of existing methods. To address this issue, for the first time, an attribute-fusion-based multimodal vehicle object detection framework is proposed: mutual distillation attribute fusion network (MDAFN). This framework employs cross-modal mutual distillation to suppress noise interference and uses an attribute decomposition approach to isolate diverse attribute features from entangled global features. Additionally, by fusing corresponding attribute features from different modalities, it achieves finer-grained feature fusion. Specifically, we first introduce a cross-modal mutual distillation module based on optimal transport theory, which leverages the complementary properties of infrared and visible images to suppress multimodal noise interference through cross-modal spatial knowledge mutual distillation. Second, we propose an attribute decomposition module that uses an attribute routing mechanism to decompose entangled multimodal global features into independent attribute features, ensuring diversity of attribute features by maximizing attribute entropy. Finally, we propose an attribute graph fusion module that independently fuses corresponding attribute features of different modalities and customizes personalized attribute graphs for each class. This operation models the associations between attribute features of various objects, thereby enhancing the model's adaptability to different object classes. Experiments on the FLIR-aligned and M³FD datasets demonstrate that MDAFN achieves state-of-the-art performance in multimodal vehicle object detection tasks. This method has broad application prospects in the scene perception tasks of autonomous driving systems.]]></description>
      <pubDate>Tue, 29 Sep 2026 16:27:50 GMT</pubDate>
      <guid>https://trid.trb.org/View/2780896</guid>
    </item>
    <item>
      <title>Ship-PFTracker: A robust ship tracking framework in short-time SAR imagery with particle filtering</title>
      <link>https://trid.trb.org/View/2728810</link>
      <description><![CDATA[Ship detection and tracking in SAR imagery are critical for maritime surveillance, vessel traffic management, and coastal security. However, accurate detection and reliable short-time tracking remain challenging due to speckle noise, sea clutter, weak target scattering, multi-scale ship variations, and complex coastal backgrounds. Existing tracking-by-detection approaches further suffer from missed detections and identity switches under small-target and noisy SAR conditions. To overcome these limitations, this project proposes Ship-PFTracker, a unified detect-then-track framework for robust SAR ship monitoring. In the detection stage, a SAR Anti-Speckle Adaptive Convolution (SAAC) module is introduced to suppress speckle noise and enhance ship structural responses. A Temporal Feature Pyramid Network (T-FPN) is designed to exploit temporal image dependencies and aggregate multi-scale contextual features, while a Cross-Scale Attention Fusion (CSAF) module refines discriminative ship representations across different resolutions. In the tracking stage, particle filtering is integrated with an adaptive IoU-confidence fusion strategy and an edge fallback mechanism to improve target association and reduce tracking loss for small ships. Experimental results on the SSTD dataset achieve 95.1% Precision, 94.3% Recall, 96.4% mAP@0.5, 81.8% MOTA, and 80.4% IDF1, offering a robust and practical solution for maritime monitoring in complex SAR environments.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:10 GMT</pubDate>
      <guid>https://trid.trb.org/View/2728810</guid>
    </item>
    <item>
      <title>Cost-Effective Radar-Based Depth Perception and Scene Interpretation</title>
      <link>https://trid.trb.org/View/2780195</link>
      <description><![CDATA[One of the current challenges for autonomous vehicles (AV) is providing provable safety in urban driving and highway driving. LIDAR provides accurate depth and shape information, but remains quite expensive. In this research we explore the viability of fusing 4D radar images with camera images as an alternative to using LIDAR. Our research aims to establish the data fusion of radar and camera images in autonomous driving as a replacement for data fusion using traditional LIDAR, radar, and camera images. Our current approach demonstrates that a trained radar detector can achieve approximately 3.5 meters of Unidirectional Chamfer Distance (UCD) against ground truth LiDAR data on a dataset recently released by Delft University of Technology (DUT) in the Netherlands, which is 28% better than the state of the art (SOTA). Our approach integrates the images obtained from deep neural network (DNN) 4D radar detectors with conventional camera RGB images. It introduces a novel pixel positional encoding algorithm inspired by Bartlett’s spectrum estimation technique. This algorithm transforms radar depth maps and RGB images into a unified pixel image subspace facilitating the learning of their potential correspondence. Our method effectively leverages high-resolution camera images to train radar depth map generative models, addressing the limitations of conventional radar detectors in complex vehicular environments while sharpening the radar output. We develop spectrum estimation algorithms tailored for radar depth maps and RGB images, a comprehensive training framework for data-driven generative models, and a camera-radar deployment scheme for AV operation.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:10 GMT</pubDate>
      <guid>https://trid.trb.org/View/2780195</guid>
    </item>
    <item>
      <title>StreetSavvi: Complete Street AI Companion for Digital Twin Driving</title>
      <link>https://trid.trb.org/View/2780190</link>
      <description><![CDATA[This research project aims at assessing the potential of Digital Twin Technology for Driving Education.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:10 GMT</pubDate>
      <guid>https://trid.trb.org/View/2780190</guid>
    </item>
    <item>
      <title>Synthesis of Capabilities, Opportunities, and Research Needs in UAS Technology for High-Precision Bridge Inspection</title>
      <link>https://trid.trb.org/View/2778182</link>
      <description><![CDATA[This report evaluates the potential role of camera-equipped unmanned aerial systems (UAS) as well as other forms of reality-mapping solutions in supporting bridge inspection practices for the Illinois Department of Transportation (IDOT). Traditional bridge inspections can require substantial labor, specialized access equipment, traffic control, and inspector exposure to hazards such as working at height, near live traffic, and over waterways. Camera-equipped UAS offer opportunities to improve visual access, reduce reliance on access-intensive equipment, minimize traffic disruption, and enhance inspection documentation. This report synthesizes federal and state case studies, regulatory requirements, emerging research, cost considerations, and IDOT bridge inventory characteristics to evaluate where UAS-based inspection can provide the greatest practical value. The findings indicate that camera-equipped UAS are most beneficial for routine visual inspection, rapid assessment, bridge deck screening, and documentation of difficult-to-access components, but are not a universal replacement for conventional methods, particularly for inspections requiring hands-on evaluation. The report also reviews emerging technologies, including autonomous navigation, computer vision-based 3D reconstruction with images and videos, artificial intelligence-assisted damage detection, automated reporting, and complementary sensing tools such as 360-degree cameras. A preliminary benefit-cost analysis and inventory-level prioritization framework are presented to identify bridge scenarios where camera-equipped UAS deployment may offer the greatest benefits in safety, accessibility, efficiency, cost-effectiveness, and inspection quality. Overall, the report recommends that reality-mapping solutions be integrated into a broader digital inspection program for IDOT, supported by targeted pilot studies, updated regulatory and cost assumptions, quality assurance protocols, and careful selection of high-value bridge inspection scenarios.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:10 GMT</pubDate>
      <guid>https://trid.trb.org/View/2778182</guid>
    </item>
    <item>
      <title>Heterogeneous bus ridership responses to micro-scale built environment: Evidence from explainable AI</title>
      <link>https://trid.trb.org/View/2781839</link>
      <description><![CDATA[Despite the inherent complexity of urban mobility, conventional methodologies often oversimplify the associations between transit demand and built environment, leaving nonlinearity and behavioral asymmetries obscured. This study integrates computer vision and explainable artificial intelligence (XGBoost-SHAP) to decode how the micro-scale built environment affects multidimensional bus ridership. Six environmental visual proxies were extracted through semantic segmentation from 3,736 street view imagery, and subsequently analyzed alongside 8.6 million ridership records over one month across 934 bus stops in Shenzhen, China. Four models distinguish temporal-behavioral patterns (weekday peak vs. weekend off-peak, boarding vs. alighting). The results reveal behavioral heterogeneity: stronger correlations were observed in boarding than alighting demand, while weekend discretionary trips exhibit greater sensitivity to micro-environmental quality than weekday mandatory trips. By quantifying nonlinear threshold effects and diminishing marginal utilities, this study develops a spatial screening matrix that supports planners in flagging environmental mismatches and advancing transit-demand-oriented, data-driven transportation planning.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:09 GMT</pubDate>
      <guid>https://trid.trb.org/View/2781839</guid>
    </item>
    <item>
      <title>Automated element-level bridge inspection from unmanned aerial vehicle imagery using 3D damage mapping and condition-state quantification</title>
      <link>https://trid.trb.org/View/2772894</link>
      <description><![CDATA[Element-level bridge inspection provides a granular representation of bridge condition but remains labor-intensive and susceptible to subjectivity and inconsistency. This paper presents an automated workflow that converts unmanned aerial vehicle imagery into element-level damage quantities and condition states consistent with the AASHTO Manual for Bridge Element Inspection. The workflow reconstructs a photogrammetric 3D model, selects a coverage-optimized subset of images, segments relevant reinforced-concrete damage classes, projects verified damage masks onto the 3D mesh, and fuses repeated multi-view observations into unified 3D damage polygons. The fused damage polygons are then associated with instance-segmented bridge elements to compute element-level damage quantities and assign condition states. Validation on a reinforced-concrete bridge demonstrates that the workflow can reproduce the dominant deterioration mechanisms and element-level condition-state distributions reported in official inspection records while substantially reducing redundant imagery. The proposed workflow provides a repeatable path from image-based damage observations to AASHTO-compatible element-level inspection outputs.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:37:08 GMT</pubDate>
      <guid>https://trid.trb.org/View/2772894</guid>
    </item>
    <item>
      <title>Degradation-Resistant Infrared-Visible Image Fusion with Auto-Generated Textual Objectives and Embedded Contrastive Learning</title>
      <link>https://trid.trb.org/View/2780910</link>
      <description><![CDATA[Infrared-visible image fusion aims to combine multi-modal image information to generate informative and robust scene representations, thereby enhancing perception capabilities and reliability in intelligent transportation systems. However, captured images often suffer from complex degradation issues, leading to low-quality source data. Existing methods are deficient in adapting to multiple degradation conditions, which limits their fusion performance. In this paper, we aim to develop a degradation-resistant image fusion method that automatically adapts to various degradations. For this purpose, we first construct an auto-generation prompt pipeline based on cascaded multi-modal and language models. It utilizes the vision-language understanding capabilities of large models to comprehensively detect degradation, then produces degradation prompts and corresponding text-based fusion objectives for each image. To resist degradations and produce the fusion results as described by fusion objectives, we next propose an embedded contrastive learning method within CLIP space to supervise the model training. This method ensures that the image fusion process is free from degradation and better aligned with the fusion objectives, which enhances the fusion model’s anti-degradation capability. Extensive experiments on public datasets validate the superiority and generalization ability of our method, and its robust degradation-adaptive capability makes it particularly suitable for complex scenes. The code is available at https://github.com/wyhlaowang/AutoFuse.]]></description>
      <pubDate>Mon, 28 Sep 2026 08:36:34 GMT</pubDate>
      <guid>https://trid.trb.org/View/2780910</guid>
    </item>
    <item>
      <title>Mapping cycling-specific infrastructure using object detection on remotely sensed images</title>
      <link>https://trid.trb.org/View/2669711</link>
      <description><![CDATA[There is growing interest for urban areas to develop road infrastructure that encourages cycling as a sustainable transport mode. However, comprehensive data on cycling infrastructure is lacking, especially for intersections. This is a crucial barrier for studying the impact of infrastructure modifications (e.g., on encouraging cycling demand or crash risks). Existing cycling infrastructure data suffers from at least one of the following deficiencies: (1) fragmentation along administrative borders, (2) link-oriented data without information on intersections, and (3) missing information on previous infrastructure changes. The contribution of this paper is threefold. First, an object detection method is utilised on aerial imagery to generate a dataset that addresses all three previously mentioned issues. For this purpose, a YOLOv11 model (a common deep learning architecture for object detection) is trained to detect ten different cycling-relevant infrastructure features, using Switzerland as a case study. Second, it is demonstrated that the method is valid by comparing a subset of the resulting dataset to an external municipality-level dataset. Third, the overall historical development of cycling-specific infrastructure in the ten largest Swiss agglomerations is presented. These results provide the basis for future analyses that depend on the knowledge of cycling infrastructure presence.]]></description>
      <pubDate>Fri, 18 Sep 2026 12:03:44 GMT</pubDate>
      <guid>https://trid.trb.org/View/2669711</guid>
    </item>
    <item>
      <title>Evaluating asphalt pavement surface temperatures using thermal satellite remote sensing</title>
      <link>https://trid.trb.org/View/2768720</link>
      <description><![CDATA[The rapid ageing of asphalt pavement surfaces due to prolonged exposure to intense thermal stresses adversely affects road performance, leading to reduced safety, lower comfort, and higher maintenance costs. Traditionally, the thermal behaviour of asphalt mixtures is investigated under controlled laboratory conditions through thermal cycling to assess performance and predict service life. Although these procedures are efficient and widely adopted, their predictive reliability diminishes under real-world conditions, primarily due to material heterogeneity and non-uniform surface temperature distributions. These challenges are intensified by climate change and the growing impact of urban heat island effects, particularly in heavily urbanised and paved environments. Satellite-based remote sensing platforms equipped with passive sensors provide historical datasets that enable long-term monitoring of land surface temperature at large scales. Multispectral imagery acquired from Landsat-8 Thermal Infrared Sensors (TIRS) and Sentinel-2 Multi-Spectral Instruments (MSI) have proven viable for analysing climate-induced changes in built and natural environments. However, for pavement-scale applications, the spatial resolution of thermal and multispectral imagery, typically ranging from 100 m to 30 m, represents a major limitation. This study proposes an innovative paradigm for high-resolution, large-scale thermal mapping of road pavements based on satellite-derived thermal anomalies. A case study at a toll highway station in the Marche Region, Italy, demonstrates the capability of the proposed methodology to accurately identify pavement surface temperature patterns.]]></description>
      <pubDate>Fri, 18 Sep 2026 11:48:31 GMT</pubDate>
      <guid>https://trid.trb.org/View/2768720</guid>
    </item>
    <item>
      <title>Design of Intelligent Traffic Light Control System Based on Vehicle Flow Monitoring Circuit</title>
      <link>https://trid.trb.org/View/2767845</link>
      <description><![CDATA[The traditional traffic signal lights adopt the switching mode of fixed countdown, which can not effectively solve the problems of low traffic efficiency and road congestion. In order to solve these above problems, the author designs an intelligent traffic light control system based on vehicle flow monitoring technology. The system as a whole is composed of functional modules such as the STM32 microcontroller module, key control module, traffic light module, LCD liquid crystal display module, Bluetooth module, and vehicle flow monitoring module. The system uses infrared photoelectric sensors and pressure sensors to monitor road vehicle flow in real time, and dynamically adjusts the countdown time of traffic lights according to road vehicle flow, so as to obtain real-time monitoring of vehicle flow to adjust the effect of traffic time, and effectively alleviate traffic congestion pressure and improve traffic guidance efficiency.]]></description>
      <pubDate>Wed, 16 Sep 2026 16:15:45 GMT</pubDate>
      <guid>https://trid.trb.org/View/2767845</guid>
    </item>
    <item>
      <title>Street-level monitoring of urban tactile paving obstructions through visual-language models and street view imagery</title>
      <link>https://trid.trb.org/View/2717792</link>
      <description><![CDATA[Tactile paving is vital infrastructure for safe mobility among 2.2 billion visually impaired individuals worldwide, but in complex urban environments it faces both static damage and dynamic encroachment. This study develops an intelligent evaluation framework that integrates visual-language models (VLMs) with pedestrian-view street imagery to assess tactile paving usability around urban metro stations. Using GPT-4o and GoPro-collected imagery, we built a three-tier risk detection system covering the tactile paving body, a 250 mm proximity zone, and the surrounding environment. The framework includes 26 structural and 24 situational indicators with differentiated risk-scoring thresholds. Based on 110 metro stations within Beijing’s Third Ring Road, we analyzed the spatial distribution of tactile paving obstructions. The 250 mm proximity zone showed the highest obstruction rate (34.46%), exceeding the tactile body (33.03%) and environment (19.79%), mainly due to spatial pressure from wall attachments, poles, and adjacent facilities. Structural obstacles reflected persistent damage and encroachment, whereas situational obstacles showed greater temporality, peak intensity, and spatial variability, especially within the proximity zone. AI evaluations closely matched expert ratings (Pearson r = 0.943), and iterative scoring reduced false positives from 54% to 11%, confirming the reliability of VLMs in complex urban contexts. Fengtai District scored poorest in both indicator categories, with Majiapu Station as a key case. We recommend introducing a “proximity buffer zone” and improving fine-scale maintenance in high-density areas. The resulting intelligent platform is scalable and transferable for nationwide monitoring and governance of accessible infrastructure.]]></description>
      <pubDate>Wed, 16 Sep 2026 11:21:26 GMT</pubDate>
      <guid>https://trid.trb.org/View/2717792</guid>
    </item>
  </channel>
</rss>