The proposed recurrent cross-frame module aggregates historical context from the previous state to enhance the coarse ground feature of each current frame and facilitates satellite candidate-region classification and hierarchical fine-grained features enable precise local offset estimation.
Abstract
Continuous and reliable localization is essential for autonomous driving. Cross-view visual localization matches ground images with satellite maps, providing complementary localization cues for pipelines that depend on Global Navigation Satellite System (GNSS) signals and high-definition (HD) maps. Most existing cross-view visual localization methods process each frame independently, leaving temporal information underused and limiting accuracy under dynamic occlusion, illumination variation, and repetitive textures. This study proposes a temporal-context-enhanced framework for cross-view sequence visual localization. The proposed recurrent cross-frame module aggregates historical context from the previous state to enhance the coarse ground feature of each current frame. These enhanced features facilitate satellite candidate-region classification, while hierarchical fine-grained features enable precise local offset estimation. On the CVIS dataset, the proposed method reduces mean localization error from 3.80 m to 1.57 m and increases R@1 m from 8.14% to 40.22%. Direct transfer to KITTI-CVL achieves a mean error of 2.61 m, with target-domain fine-tuning further reducing the mean error to 2.27 m. Zero-shot field experiments on a real-world vehicle achieve a mean error of 2.84 m and R@5 m of 96.86%. These results demonstrate that temporal context enhancement significantly improves cross-view localization accuracy and supports robust deployment on public benchmarks and real-world roads.
Cross-view geo-localization (CVGL) aims to match images captured from different viewpoints, such as drone and satellite imagery. Existing methods primarily focus on single-image matching, overlooking the potential of leveraging temporal information from drone image sequences. To address this, we propose Spatio-Temporal Alignment and Refinement (STAR), a novel framework that effectively leverages the spatio-temporal dependencies of drone sequences for effective matching with satellite images. Notably, this is among the first attempts to explore drone-sequence-based matching in CVGL, further improving retrieval performance through spatio-temporal consistency. Specifically, we design the Video Vision Transformer (ViViT) Sequence Alignment Module (VSAM) to effectively extract and align sequential features and introduce a Pseudo-Temporal Compensation (PTC) strategy to prevent the temporal modeling branch from degenerating when processing static satellite inputs, maintaining architectural symmetry across dynamic and static views. In addition, we design the Dynamic Background Partitioning Module (DBPM) to adaptively segment features and improve foreground–background distinction. Extensive experiments on the University-1652 and SUES-200 datasets demonstrate that our method significantly outperforms state-of-the-art approaches, highlighting the benefits of incorporating sequence information in CVGL.
Ziqian Mo, Yuxi Sun, Sen Jia et al.· IEEE Transactions on Geoscie...· 0 citations
With rising performance demands for autonomous driving perception systems, bird's-eye-view (BEV) perception that combines heterogeneous sensor data and temporal information has become a key research focus. Recent studies have advanced either multimodal temporal detection through BEV-level feature aggregation or sparse object-centric temporal modeling. However, integrating these two directions while preserving modality-specific temporal states remains less explored for LiDAR-camera detection. In this work, we present TempoCross, a 3D detection method based on instance-aware sparse representations for multimodal temporal fusion. TempoCross encodes features from different timestamps and sensor modalities in a unified instance space, enabling adaptive extraction of critical target information from modality- and time-specific states. Initially, both branches perform a preliminary cross-modal fusion to generate queries for the current frame. In the motion compensation module, a hybrid motion modeling strategy reduces alignment discrepancies between historical and current instances caused by complex object motion. This strategy combines explicit rigid-body transformations with implicit learnable deformation residuals, improving both accuracy and robustness in cross-frame instance association. Next, temporal-aware enhancement integrates the initial queries and current-frame features with motion-compensated historical instances. Finally, a lightweight cross-attention fuses current and historical instances from both branches. This formulation reduces the reliance on repeatedly propagating full-scene fused BEV features and concentrates temporal interaction on target-related instance states. On the nuScenes test set, TempoCross achieves 74.1% mAP and 75.7% NDS, outperforming mainstream baseline detectors. The results support the effectiveness of combining LiDAR-camera fusion with instance-aware sparse temporal modeling.
Accurate and reliable pose information with respect to a reference frame is increasingly demanded across applications such as autonomous navigation, surveying, robotics, and augmented and mixed reality. Visual localization can serve as a complementary positioning modality to GNSS, whose applicability and accuracy are often limited. Yet, the accuracy potential of visual localization has not been systematically investigated against survey-grade demands. This is mainly due to the lack of publicly available, large-scale outdoor datasets with ground-truth poses in the sub-centimeter range. In this work, we address both gaps. We introduce a scalable visual localization pipeline that employs precisely georeferenced, high-resolution street-level imagery directly as the scene representation. It combines prior-guided reference candidate selection with on-the-fly local Structure-from-Motion reconstruction and PnP-based pose estimation. We further present the FHNW Muttenz dataset, a real-world dataset covering a contiguous 10 km street network mapped in two mobile mapping campaigns approximately 1.5 years apart. It consists of high-resolution reference imagery and query sequences acquired by four different cameras across five representative scenes. All images are precisely co-registered, yielding 6-DoF ground-truth poses in the sub-centimeter range. Using this dataset, we evaluate the accuracy potential of visual localization. Our experiments demonstrate median pose accuracies in the range of 1-5 cm for translation and 0.05-0.1{\deg} for rotation, reaching as low as 1 cm and 0.03{\deg} under favorable conditions. These results show that visual localization can complement survey-grade GNSS positioning, paving the way for 3D geospatial data acquisition using consumer devices and fully automated georeferencing approaches. The dataset is publicly available at: https://fhnw-muttenz-vl-dataset.github.io/.
Jonas Meyer, S. Nebiker, P. Theiler et al.· 0 citations
Visual localization is vital for autonomous systems but remains challenging under dynamic conditions. Transformers offer strong temporal modeling at quadratic cost, while CNNs are efficient yet limited in long-range dependencies. Existing methods also lack robustness to illumination, weather, and seasonal changes, constraining real-world applicability. To address this, this paper proposes AdapseqNet, a dual-branch architecture that integrates stabilized state-space modeling with differential temporal enhancement. First, a stabilized state-space formulation featuring Lyapunov-constrained parameterization and adaptive discretization is proposed, ensuring asymptotic stability and linear computational complexity for reliable processing of extended sequences. Second, a selective Mamba architecture is developed to combine temporal-state modeling with content-aware gating, enabling adaptive feature selection that emphasizes discriminative cues while suppressing redundancy. Third, a differential enhancement module is designed to extract motion-invariant representations through symmetric temporal differencing and LSTM-based refinement, enhancing resilience to appearance variations caused by lighting, weather, and seasonal changes. Beyond architectural design, multi-scale feature fusion and output distribution control are incorporated to optimize representation quality and ensure consistency for similarity-based retrieval. Extensive experiments on multiple benchmarks demonstrate that AdapseqNet achieves a better localization accuracy across diverse and challenging conditions. Note to Practitioners—Visual localization is crucial for autonomous robots but often fails under varying lighting, weather, or seasonal conditions. We propose a dual-path approach: one path captures long-term patterns using control-inspired stable modeling, while the other extracts motion cues that remain consistent despite appearance changes. This combination enables accurate place recognition even in extreme environments. Our system operates efficiently on standard hardware and was tested on an indoor robot, achieving centimeter-level accuracy. This approach can enhance existing navigation systems without requiring additional sensors. Future work will focus on real-time optimization for outdoor deployment.
Zhenyu Li, Tianyi Shang· IEEE Transactions on Automat...· 0 citations
Accurate traffic scene recognition serves as a critical foundation for decision-making and safe driving in autonomous driving and intelligent transportation systems. Existing methods mostly rely on single visual data vulnerable to environmental variations, or vision-LiDAR fusion schemes with insufficient capacity to represent road topology and traffic semantic information, limiting recognition accuracy and robustness. To address these limitations, this paper proposes an end-to-end traffic scene recognition network based on the fusion of monocular camera images and corresponding road map top-down view data. We design a learnable cross-view spatial alignment module to eliminate perspective discrepancy, and a bidirectional cross-attention fusion module to enable deep bidirectional interaction between visual semantic and map topology features. Experiments on a self-built dataset covering five typical traffic scenes show that the proposed method achieves an overall recognition accuracy of 92.6%, outperforming the best single-input baseline Swin-Tiny by 3.3%. Ablation studies further validate the effectiveness of each core module.
ZhenYu Cheng, Haoyu Kon· International Conference on...· 0 citations
Multiview geo-localization utilizing drone and satellite imagery offers a reliable alternative to GPS-based positioning in challenging environments such as urban canyons and electromagnetically degraded areas. A key challenge in this task stems from the spatial misalignment between drone-view and satellite-view images, particularly when target buildings appear off-center due to varying flight attitudes and environmental disturbances. Existing methods, which often rely on implicit center-aligned assumptions, show limited robustness under such offset conditions. To address this limitation, we construct a novel drone-view offset dataset named Offset-1652 by applying controlled translational transformations to the benchmark dataset, simulating realistic displacement scenarios along horizontal, vertical, and diagonal directions. Furthermore, we propose a large-kernel perceptual attention network (LK-PAN) that employs large-kernel depthwise convolutions to expand the receptive fields, thereby capturing global contextual information even for off-center targets. A symmetric InfoNCE loss is introduced to enhance cross-modal feature alignment and improve discrimination of hard negative samples. Comprehensive experiments conducted on both established benchmark datasets and the newly developed Offset-1652 and Offset-1652-MIX datasets demonstrate that the proposed method significantly outperforms existing approaches under a wide range of offset conditions. These results confirm its superior robustness and generalization capability for practical multiview geo-localization in complex operational environments. The source code and datasets are available at https://github.com/HAORANJY/LK-PAN-main
Bangyong Sun, Mian Li, Weifeng Wang et al.· IEEE Transactions on Geoscie...· 0 citations