LiDAR super-resolution enables high-quality 3D perception from cost-effective, low-resolution sensors. While transformer-based approaches like TULIP show promise, they remain limited to spatial-domain processing with restricted receptive fields. We introduce FLASH, a novel framework featuring two key innovations: (i) Frequency-Aware Window Attention combining local spatial attention with global frequency-domain analysis via Fast Fourier Transform, and (ii) Adaptive Multi-Scale Fusion replacing conventional skip connections with learned position-specific feature aggregation enhanced by Convolutional Block Attention Module. Experiments on KITTI demonstrate state-of-the-art performance across all metrics, outperforming even TULIP with Monte Carlo Dropout while maintaining single-pass efficiency at 66 FPS. Our dual-domain approach effectively handles uncertainty through architectural design, making it practical for real-time autonomous systems.

Visual Place Recognition (VPR) aims to match a query image to reference images of the same place in a large-scale database. Recent state-of-the-art methods employ Vision Transformers (ViTs) as backbone foundation models to extract patch-level features that are robust to viewpoint, illumination, and seasonal variations, which are then aggregated into a compact global descriptor for retrieval. Most existing aggregation methods uniformly pool patch tokens into learned clusters, despite the fact that different clusters often encode distinct spatial or semantic patterns and contribute unequally to VPR performance. To address this limitation, we propose Weighted Aggregated Descriptor (WeiAD), which assigns weights to clusters during aggregation, producing more discriminative global representations. Beyond accuracy, retrieval latency is a critical concern for large-scale deployments and resource-constrained edge devices. Prior work mainly reduces latency by compressing global descriptors, while overlooking the cost of feature extraction, an issue exacerbated by ViT-based backbones. We therefore introduce WeiToP, a VPR-oriented token pruning framework that reduces feature extraction cost via self-distillation, where aggregation-induced token importance supervises a lightweight pruning module attached to an early transformer layer, enabling inference-time token pruning. After a single joint training phase, WeiToP enables plug-and-play token pruning at inference time, allowing flexible and on-demand control over the accuracy-efficiency trade-off without additional training. Moreover, WeiToP outperforms existing token pruning methods adapted from general vision tasks.

Accurate perception is critical for vehicle safety, with LiDAR as a key enabler in autonomous driving. To ensure robust performance across environments, sensor types, and weather conditions without costly re-annotation, domain generalization in LiDAR-based 3D semantic segmentation is essential. However, LiDAR annotations are often noisy due to sensor imperfections, occlusions, and human errors. Such noise degrades segmentation accuracy and is further amplified under domain shifts, threatening system reliability. While noisy-label learning is well-studied in images, its extension to 3D LiDAR segmentation under domain generalization remains largely unexplored, as the sparse and irregular structure of point clouds limits direct use of 2D methods. To address this gap, we introduce the novel task Domain Generalization for LiDAR Semantic Segmentation under Noisy Labels (DGLSS-NL) and establish the first benchmark by adapting three representative noisy-label learning strategies from image classification to 3D segmentation. However, we find that existing noisy-label learning approaches adapt poorly to LiDAR data. We therefore propose DuNe, a dual-view framework with strong and weak branches that enforce feature-level consistency and apply cross-entropy loss based on confidence-aware filtering of predictions. Our approach shows state-of-the-art performance by achieving 56.86% mIoU on SemanticKITTI, 42.28% on nuScenes, and 52.58% on SemanticPOSS under 10% symmetric label noise, with an overall Arithmetic Mean (AM) of 49.57% and Harmonic Mean (HM) of 48.50%, thereby demonstrating robust domain generalization in DGLSS-NL tasks. The code is available at https://github.com/MKong17/DGLSS-NL.git

Cross-view Video Geo-localization (CVG) aims to localize ground-view videos by retrieving their corresponding geo-tagged aerial images. However, CVG approaches rely on fixed-length inputs and post-hoc refinement, hindering online-oriented localization under partial or dynamic observations. In this work, we formulate Progressive Cross-view Video Geo-localization (PCVG) as a deployment-oriented extension and evaluation protocol of CVG, enabling localization under varying temporal budgets, prefix-based inference, random-start evaluation, and long-range localization with interruptions. To explore PCVG, we introduce X^2Localizer, a cross-grained alignment framework that jointly supervises global prefix-to-aerial retrieval and token-aggregated frame–aerial-tile matching with a budget-dependent asymmetric objective. Furthermore, we introduce a Sliding-Window Re-Localization (SWRL) strategy that dynamically refreshes candidate regions for failure recovery and long-range deployment without full-sequence reprocessing. Extensive experiments show that X^2Localizer preserves conventional full-video performance, with marginal gains of +0.1 Recall@1 and +0.3 Recall@10, while substantially improving early localization. In the challenging single-frame setting, X^2Localizer improves coarse retrieval by +4.7 Recall@1 and +11.5 Recall@10 over the previous state-of-the-art method. With SWRL, our approach further enables robust progressive localization under random-start and long-distance scenarios, narrowing the gap between benchmark evaluation and real-world deployment.

Visual Place Recognition (VPR) is a fundamental task in robotics and computer vision, enabling systems to identify locations seen in the past using visual information. Previous state-of-the-art approach - SegVLAD - focuses on encoding and retrieving semantically meaningful supersegment representations of images to significantly enhance recognition recall rates. However, we find that they struggle to cope with significant variations in viewpoint and scale, as well as scenes with sparse or limited information. Furthermore, these semantic-driven supersegment representations often exclude semantically meaningless yet valuable pixel information. In this work, we present Sel-V and MuSSel-V, two efficient variants within the segment-level VPR paradigm that replace heavy and fragmented supersegments with lightweight, visually compact and complete dilated superpixels for local feature aggregation. The use of superpixels preserves pixel-level details while reducing computational overhead. A multi-scale extension further enhances robustness to viewpoint and scale changes. Comprehensive experiments on twelve public benchmarks show that our approach achieves a better trade-off between accuracy and efficiency than existing segment-based methods. These results demonstrate that lightweight, non-semantic segmentation can serve as an effective alternative for high-performance, resource efficient visual place recognition in robotics.

LiDAR sensors are often considered essential for autonomous driving, but high-resolution sensors remain expensive while affordable low-resolution sensors produce sparse point clouds that miss critical details. LiDAR super-resolution addresses this challenge by using deep learning to enhance sparse point clouds, bridging the gap between different sensor types and enabling cross-sensor compatibility in real-world deployments. This paper presents the first comprehensive survey of LiDAR super-resolution methods for autonomous driving. Despite the importance of practical deployment, no systematic review has been conducted until now. We organize existing approaches into four categories: CNN-based architectures, model-based deep unrolling, implicit representation methods, and Transformer and Mamba-based approaches. We establish fundamental concepts including data representations, problem formulation, benchmark datasets and evaluation metrics. Current trends include the adoption of range image representation for efficient processing, extreme model compression and the development of resolution-flexible architectures. Recent research prioritizes real-time inference and cross-sensor generalization for practical deployment. We conclude by identifying open challenges and future research directions for advancing LiDAR super-resolution technology.

It is essential to detect and segment cracks in various infrastructures, such as roads and buildings, to ensure safety, longevity, and cost-effective maintenance. Despite deep learning advancements, precise crack detection across diverse conditions remains challenging. This paper introduces Hybrid-Segmentor, a deep learning model combining Convolutional Neural Networks-based and Transformer-based architectures to extract both fine-grained local features and global crack patterns, significantly enhancing crack detection for improved infrastructure maintenance. Hybrid-Segmentor, trained on a large custom dataset created by merging multiple open-source datasets, can accurately detect cracks on different types of surfaces, crack shapes, and sizes. The model demonstrates robustness and versatility by accurately detecting discontinuities, vague cracks, non- crack regions within crack areas, blurred images, and complex crack contours. Furthermore, when compared against other recent models for crack segmentation, the proposed model achieves state-of-the-art performance, significantly outperforming them across five key metrics: accuracy (0.971), precision (0.807), recall (0.756), F1-score (0.774), and IoU (0.631).

In this work we investigate the problem of measuring a checkerboard target’s centre in an 3D point cloud. This is an important problem which has applications in registration, long term monitoring and linking to other sensor systems. We use a 3D template matching approach based on the coloured ICP algorithm to solve the problem. We tackle the problem under the additional constraints that we assume no structure in the 3D data in order to be able to handle unordered point clouds. This gives us the capability to process data from the new generation of low-cost LIDAR sensors. This category of sensors also suffers from increased noise in range and reflectivity measurement. We provide extensive simulation results using synthetic data to capture the potential of the approach. We then give the detailed steps for handling real sensor data.

LiDAR super-resolution addresses the challenge of achieving high-quality 3D perception from cost-effective, low-resolution sensors. While recent transformer-based approaches like TULIP show promise, they remain limited to spatial-domain processing with restricted receptive fields. We introduce FLASH (Frequency-aware LiDAR Adaptive Super-resolution with Hierarchical fusion), a novel framework that overcomes these limitations through dual-domain processing. FLASH integrates two key innovations: (i) Frequency-Aware Window Attention that combines local spatial attention with global frequency-domain analysis via FFT, capturing both fine-grained geometry and periodic scanning patterns at log-linear complexity. (ii) Adaptive Multi-Scale Fusion that replaces conventional skip connections with learned position-specific feature aggregation, enhanced by CBAM attention for dynamic feature selection. Extensive experiments on KITTI demonstrate that FLASH achieves state-of-the-art performance across all evaluation metrics, surpassing even uncertainty-enhanced baselines that require multiple forward passes. Notably, FLASH outperforms TULIP with Monte Carlo Dropout while maintaining single-pass efficiency, which enables real-time deployment. The consistent superiority across all distance ranges validates that our dual-domain approach effectively handles uncertainty through architectural design rather than computationally expensive stochastic inference, making it practical for autonomous systems.

A building’s age of construction is crucial for supporting many geospatial applications. Much current research focuses on estimating building age from facade images using deep learning. However, building an accurate deep learning model requires a considerable amount of labelled training data, and the trained models often have geographical constraints. Recently, large pre-trained vision language models (VLMs) such as GPT-4 Vision, which demonstrate significant generalisation capabilities, have emerged as potential training-free tools for dealing with specific vision tasks, but their applicability and reliability for building information remain unexplored. In this study, a zero-shot building age classifier for facade images is developed using prompts that include logical instructions. Taking London as a test case, we introduce a new dataset, FI-London, comprising facade images and building age epochs. Although the training-free classifier achieved a modest accuracy of 39.69%, the mean absolute error of 0.85 decades indicates that the model can predict building age epochs successfully albeit with a small bias. The ensuing discussion reveals that the classifier struggles to predict the age of very old buildings and is challenged by fine-grained predictions within 2 decades. Overall, the classifier utilising GPT-4 Vision is capable of predicting the rough age epoch of a building from a single facade image without any training. The code and dataset are available at https://zichaozeng.github.io/ba_classifier.

Recent advances have demonstrated that Language Vision Models (LVMs) surpass the existing State-of-the-Art (SOTA) in two-dimensional (2D) computer vision tasks, motivating attempts to apply LVMs to three-dimensional (3D) data. While LVMs are efficient and effective in addressing various downstream 2D vision tasks without training, they face significant challenges when it comes to point clouds, a representative format for representing 3D data. It is more difficult to extract features from 3D data and there are challenges due to large data sizes and the cost of the collection and labelling, resulting in a notably limited availability of datasets. Moreover, constructing LVMs for point clouds is even more challenging due to the requirements for large amounts of data and training time. To address these issues, our research aims to 1) apply the Grounded SAM through Spherical Projection to transfer 3D to 2D, and 2) experiment with synthetic data to evaluate its effectiveness in bridging the gap between synthetic and real-world data domains. Our approach exhibited high performance with an accuracy of 0.96, an IoU of 0.85, precision of 0.92, recall of 0.91, and an F1 score of 0.92, confirming its potential. However, challenges such as occlusion problems and pixel-level overlaps of multi-label points during spherical image generation remain to be addressed in future studies.

This study investigates the efficacy of an open vocabulary, multi-modal, foundation model for the semantic segmentation of images from complex urban street scenes. Unlike traditional models reliant on predefined category sets, Grounded SAM uses arbitrary textual inputs for category definition, offering enhanced flexibility and adaptability. The model’s performance was evaluated across single and multiple category tasks using the benchmark datasets Cityscapes, BDD100K, GTA5, and KITTI. The study focused on the impact of textual input refinement and the challenges of classifying visually similar categories. Results indicate strong performance in single-category segmentation but highlighted difficulties in multi-category scenarios, particularly with categories bearing close textual or visual resemblances. Adjustments in textual prompts significantly improved detection accuracy, though challenges persisted in distinguishing between visually similar objects such as buses and trains. Comparative analysis with state-of-the-art models revealed Grounded SAM’s competitive performance, particularly notable given its direct inference capability without extensive dataset-specific training. This feature is advantageous for resource-limited applications. The study concludes that while open vocabulary models such as Grounded SAM mark a significant advancement in semantic segmentation, further improvements in integrating image and text processing are essential for better performance in complex scenarios.

Terrestrial LiDAR is an established method for 3D data acquisition in close-range applications. A new category of low-cost LiDAR sensors for autonomous driving has become available with similar specifications. However, these new sensors lack the guarantee of survey-grade performance. Initial experiments have broadly confirmed the specification of one such low-cost sensor but have also raised issues with the radiometric behaviour. This study investigates through practical experiments how the intensity information of the Livox-Mid40 laser scanner is influenced by time, reflectance, distance and angle of incidence. The quantitative analysis of the experiments shows an expected relationship between surface reflectance and recorded intensity. However, the result indicate that intensity can significantly influence the distance measurement at very close ranges.

Traditional safety analysis methods based on historical crash data and simulation models have limitations in capturing real-world driving scenarios. In this experiment, panoramic videos recorded from a motorcyclist’s helmet in Bangkok, Thailand, were narrated using an image-to-text model and then put into a Large Language Model (LLM) to identify potential hazards and assess crash risks. The framework can assess static and moving objects with the potential for early warning and incident analysis. However, the limitations of the existing image-to-text model cause its inability to handle panoramic images effectively. 2012 ACM Subject Classification Information systems → Geographic information systems; Computing methodologies → Scene understanding

The poor documentation of subsurface utility data is a common problem in many cities, exposing field engineers to risks of utility strike. This paper investigates the use of consumer-grade cameras to improve operational efficiency on construction sites and explores different imaging networks to optimize photogrammetric processing for low-cost subsurface utility surveys. Results from the first part of the study demonstrated the potential of consumer-grade cameras as a photogrammetric utility data acquisition tool. However, statistical insights from the photogrammetric calibration show that caution needs to be taken about the camera types particularly for lens calibration. Results from the second part of the study were recommended as easy-to-understand guidelines for image acquisition at trenches and supported the planning of photogrammetric measurements in the field.

Deep Learning (DL) semantic image segmentation is a technique used in several fields of research. The present paper analyses semantic crack segmentation as a case study to review the up to date research on semantic segmentation in the presence of fine structures and the effectiveness of established approaches to address the inherent class imbalance issue. The established UNet architecture is tested against networks consisting exclusively of stacked convolution without pooling layers (straight networks), with regard to the resolution of their segmentation results. Dice and Focal losses are also compared against each other to evaluate their effectiveness on highly imbalanced data. With the same aim, dropout and data augmentation approaches are tested, as additional regularizing mechanisms, to address the uneven distribution of the dataset. The experiments show that the good selection of the loss function has more impact in handling the class imbalance and boosting the detection performance than all the other regularizers with regards to segmentation resolution. Moreover, UNet, the architecture considered as reference, clearly outperforms the networks with no pooling layers both in performance and training time. The authors argue that UNet architectures, compared to the networks with no pooling layers, achieve high detection performance at a very low cost in terms of training time. Therefore, the authors consider such architecture as the state of the art for semantic segmentation of cracks. On the other hand, once computational cost is not an issue anymore thanks to constant improvements of technology, the application of networks without pooling layers might become attractive again because of their simplicity of and high performance.

This paper proposes a semantic segmentation pipeline for terrestrial laser scanning data. We achieve this by combining co-registered RGB and 3D point cloud information. Semantic segmentation is performed by applying a pre-trained off-the-shelf 2D convolutional neural network over a set of projected images extracted from a panoramic photograph. This allows the network to exploit the visual image features that are learnt in a state-of-the-art segmentation models trained on very large datasets. The study focuses on the adoption of the spherical information from the laser capture and assessing the results using image classification metrics. The obtained results demonstrate that the approach is a promising alternative for asset identification in laser scanning data. We demonstrate comparable performance with spherical machine learning frameworks, however, avoid both the labelling and training efforts required with such approaches.

In this paper we set out to solve the task of 6-DOF 3D object detection from 2D images, where the only supervision is a geometric representation of the objects we aim to find. In doing so, we remove the need for 6-DOF labels (i.e., position, orientation etc.), allowing our network to be trained on unlabeled images in a self-supervised manner. We achieve this through a neural network which learns an explicit scene parameterization which is subsequently passed into a differentiable renderer. We analyze why analysis-by-synthesis-like losses for supervision of 3D scene structure using differentiable rendering is not practical, as it almost always gets stuck in local minima of visual ambiguities. This can be overcome by a novel form of training, where an additional network is employed to steer the optimization itself to explore the entire parameter space i.e., to be curious, and hence, to resolve those ambiguities and find workable minima.

Massive semantically labeled datasets are readily available for 2D images, however, are much harder to achieve for 3D scenes. Objects in 3D repositories like ShapeNet are labeled, but regrettably only in isolation, so without context. 3D scenes can be acquired by range scanners on city-level scale, but much fewer with semantic labels. Addressing this disparity, we introduce a new optimization procedure, which allows training for 3D detection with raw 3D scans while using as little as 5% of the object labels and still achieve comparable performance. Our optimization uses two networks. A scene network maps an entire 3D scene to a set of 3D object centers. As we assume the scene not to be labeled by centers, no classic loss, such as Chamfer can be used to train it. Instead, we use another network to emulate the loss. This loss network is trained on a small labeled subset and maps a non-centered 3D object in the presence of distractions to its own center. This function is very similar – and hence can be used instead of – the gradient the supervised loss would provide. Our evaluation documents competitive fidelity at a much lower level of supervision, respectively higher quality at comparable supervision. Supplementary material can be found at: dgriffiths3.github.io .