Temporal Vector Map Perception with Enhanced Query and Map Prior
LIN Yutong1, YUAN Wei2,3,4, ZHUANG Hanyang5, QIAN Yeqiang1,6
1. School of Automation and Intelligent Sensing, Shanghai Jiao Tong University, Shanghai 200240; 2. School of Geospatial Artificial Intelligence, East China Normal University, Shanghai 200241; 3. Key Laboratory of Geographic Information Science, Ministry of Education, East China Normal University, Shanghai 200241; 4. Key Laboratory of Spatial-Temporal Big Data Analysis and App-lication of Natural Resources in Megacities, Ministry of Na-tural Resources, East China Normal University, Shanghai 200241; 5. Global College, Shanghai Jiao Tong University, Shanghai 200240; 6. Key Laboratory of System Control and Information Processing, Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240
Abstract：Online construction of vectorized high-definition(HD) map is one of the critical tasks in autonomous driving system. However, the prediction performance of the existing methods is reduced under complex conditions, such as occlusion and low illumination. To address this issue, a method for temporal vector map perception with enhanced query and map prior(QPMap) is proposed. First, a Transformer-based standard definition map enhancement module and a rasterized high definition map initialization module are designed to enhance environmental understanding in complex scenarios. Second, an instance-level query encoding strategy is designed. The detection results from previous frames are introduced as temporal query priors into the current perception process to improve prediction stability and temporal consistency. Finally, a multi-scale decoder and a direction vector loss function are utilized to refine query representations. Thus, vectorized map elements are predicted accurately. A real-world dataset is constructed and a fisheye camera adaptation strategy is developed to validate the applicability of QPMap in actual autonomous driving scenarios. Experiments on multiple public datasets and a self-collected dataset demonstrate that QPMap achieves superior map construction performance while maintaining good real-time performance. Moreover, QPMap effectively improves the accuracy and robustness of online HD map construction.
[1] LIAO B C, CHEN S Y, WANG X G, et al. MapTR: Structured Modeling and Learning for Online Vectorized HD Map Construction[C/OL].[2026-02-16]. https://arxiv.org/pdf/2208.14437. [2] LIAO B C, CHEN S Y, ZHANG Y C, et al. MapTRv2: An End-to-End Framework for Online Vectorized HD Map Construction. International Journal of Computer Vision, 2025, 133(3): 1352-1374. [3] CAESAR H, BANKITI V, LANG A H, et al. nuScenes: A Multimodal Dataset for Autonomous Driving // Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2020: 11618-11628. [4] WILSON B, QI W, AGARWAL T, et al. Argoverse 2: Next Gene-ration Datasets for Self-Driving Perception and Forecasting[C/OL].[2026-02-16]. https://arxiv.org/pdf/2301.00493. [5] JIANG B, CHEN S Y, XU Q, et al. VAD: Vectorized Scene Re-presentation for Efficient Autonomous Driving // Proc of the IEEE/CVF International Conference on Computer Vision. Washington, USA: IEEE, 2023: 8306-8316. [6] JIANG B, CHEN S Y, GAO H, et al. VADV2: End-to-End Vectorized Autonomous Driving via Probabilistic Planning[C/OL].[2026-02-16]. https://arxiv.org/pdf/2402.13243. [7] HU Y H, YANG J Z, CHEN L, et al. Planning-Oriented Autonomous Driving // Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2023: 17853-17862. [8] LI Y H, GE Z, YU G Y, et al. BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection. Proceedings of the AAAI Conference on Artificial Intelligence, 2023, 37(2): 1477-1485. [9] LIU Y C, YUAN T Y, WANG Y, et al. VectorMapNet: End-to-End Vectorized HD Map Learning // Proc of the 40th International Conference on Machine Learning. San Diego, USA: JMLR, 2023: 22352-22369. [10] CARION N, MASSA F, SYNNAEVE G, et al. End-to-End Object Detection with Transformers // Proc of the 16th European Confe-rence on Computer Vision. Berlin, Germany: Springer, 2020: 213-229. [11] QIAO L M, DING W J, QIU X, et al. End-to-End Vectorized HD-Map Construction with Piecewise Bézier Curve // Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2023: 13218-13228. [12] YUAN T Y, LIU Y C, WANG Y, et al. StreamMapNet: Streaming Mapping Network for Vectorized Online HD Map Construction // Proc of the IEEE/CVF Winter Conference on Applications of Computer Vision. Washington, USA: IEEE, 2024: 7341-7350. [13] CHEN J C, WU Y F, TAN J Q, et al. MapTracker: Tracking with Strided Memory Fusion for Consistent Vector HD Mapping[C/OL].[2026-02-16]. https://arxiv.org/pdf/2403.15951v1. [14] LIU Z H, ZHANG X Y, LIU G W, et al. Leveraging Enhanced Queries of Point Sets for Vectorized Map Construction // Proc of the 18th European Conference on Computer Vision. Berlin, Germany: Springer, 2024: 461-477. [15] WANG R X, LU X, LIU X Y, et al. PriorMapNet: Enhancing Online Vectorized HD Map Construction with Priors[C/OL].[2026-02-16]. https://arxiv.org/pdf/2408.08802v1. [16] ZHANG Y R, QIAN Y Q, MENG C, et al. Local Vectorized High Definition Map Construction for Autonomous Driving: A Comprehensive Review. IEEE Transactions on Intelligent Transportation Systems, 2025, 26(12): 21502-21525. [17] LUO K Z, WENG X S, WANG Y, et al. Augmenting Lane Perception and Topology Understanding with Standard Definition Navigation Maps // Proc of the IEEE International Conference on Robotics and Automation. Washington, USA: IEEE, 2024: 4029-4035. [18] JIANG Z, ZHU Z X, LI P F, et al. P-MapNet: Far-Seeing Map Generator Enhanced by both SDMap and HDMap Priors. IEEE Robotics and Automation Letters, 2024, 9(10): 8539-8546. [19] GAO W J, FU J W, SHEN Y Q, et al. Complementing Onboard Sensors with Satellite Maps: A New Perspective for HD Map Construction // Proc of the IEEE International Conference on Robotics and Automation. Washington, USA: IEEE, 2024: 11103-11109. [20] SUN R, YANG L, LINGRAND D, et al. Mind the Map! Accoun-ting for Existing Maps Information When Estimating Online HDMaps from Sensors // Proc of the IEEE/CVF Winter Conference on Applications of Computer Vision. Washington, USA: IEEE, 2025: 1671-1681. [21] ZHANG X Y, LIU G W, LIU Z H, et al. Enhancing Vectorized Map Perception with Historical Rasterized Maps // Proc of the 18th European Conference on Computer Vision. Berlin, Germany: Springer, 2024: 422-439. [22] XIONG X, LIU Y C, YUAN T Y, et al. Neural Map Prior for Autonomous Driving // Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2023: 17535-17544. [23] LI Z Q, WANG W H, LI H Y, et al. BEVFormer: Learning Bird's-Eye-View Representation from Multi-camera Images via Spatiotemporal Transformers // Proc of the 17th European Conference on Computer Vision. Berlin, Germany: Springer, 2022: 1-18. [24] WANG S, JIA F, MAO W X, et al. Stream Query Denoising for Vectorized HD-Map Construction // Proc of the 18th European Conference on Computer Vision. Berlin, Germany: Springer, 2024: 203-220. [25] HAKLAY M, WEBER P.OpenStreetMap: User-Generated Street Maps. IEEE Pervasive Computing, 2008, 7(4): 12-18. [26] PHILION J, FIDLER S. Lift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D // Proc of the 16th European Conference on Computer Vision. Berlin, Germany: Springer, 2020: 194-210. [27] HE K M, ZHANG X Y, REN S Q, et al. Deep Residual Learning for Image Recognition // Proc of the IEEE Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2016: 770-778. [28] WOO S, PARK J, LEE J Y, et al. CBAM: Convolutional Block Attention Module // Proc of the 15th European Conference on Computer Vision. Berlin, Germany: Springer, 2018: 3-19. [29] 黄峰,刘鸿伟,沈英,等.基于多尺度空间自适应注意力网络的轻量级图像超分辨率方法.模式识别与人工智能, 2025, 38(1): 36-50. (HUANG F, LIU H W, SHEN Y, et al. Lightweight Image Super-Resolution Reconstruction Method Based on Multi-scale Spatial Adaptive Attention Network. Pattern Recognition and Artificial Intelligence, 2025, 38(1): 36-50.) [30] MENG D P, CHEN X K, FAN Z J, et al. Conditional DETR for Fast Training Convergence // Proc of the IEEE/CVF International Conference on Computer Vision. Washington, USA: IEEE, 2021: 3631-3640. [31] HUANG J J, HUANG G.BEVPoolv2: A Cutting-Edge Implementation of Bevdet Toward Deployment[C/OL].[2026-02-16].https://arxiv.org/pdf/2211.17111v1. [32] DING W J, QIAO L M, QIU X, et al. PivotNet: Vectorized Pivot Learning for End-to-End HD Map Construction // Proc of the IEEE/CVF International Conference on Computer Vision. Wa-shington, USA: IEEE, 2023: 3649-3659. [33] ZHANG G J, LIN J H, WU S, et al. Online Map Vectorization for Autonomous Driving: A Rasterization Perspective // Proc of the 37th International Conference on Neural Information Processing Systems. Cambridge, USA: MIT Press, 2023: 31865-31877. [34] LIU X L, WANG S, LI W T, et al. MGMap: Mask-Guided Lear-ning for Online Vectorized HD Map Construction // Proc of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA: IEEE, 2024: 14812-14821. [35] CHOI S, KIM J, SHIN H, et al. Mask2Map: Vectorized HD Map Construction Using Bird's Eye View Segmentation Masks // Proc of the 18th European Conference on Computer Vision. Berlin, Germany: Springer, 2024: 19-36.
2026, Vol. 39 
