来源:时空探索之旅
原文链接:https://mp.weixin.qq.com/s/Ruuwr2CLXZhRGQJaM0-OhQ
ECCV 2026将在2026年9月8日至12日于瑞典马尔默(Malmö, Sweden)举行。
本文总结了ECCV 2026上有关时空时序的相关论文。
时空时序Topic:交通模拟,自动驾驶,世界模型,多模态大模型,空间推理,天气预报等。
ECCV2026论文列表:https://eccv.ecva.net/virtual/2026/papers.html
2. ECoSim: Data Efficient Fine-Tuning for Controllable Traffic Simulation 3. RoadBench: Benchmarking MLLMs on Fine-Grained Spatial Understanding and Reasoning under Urban Road Scenarios 4. OpenEarthAgent: A Unified Framework for Tool-Augmented Geospatial Agents 5. UrbanAlign: Post-hoc Semantic Calibration for VLM-Human Preference Alignment 6. Driver-WM: A Driver-Centric Traffic-Conditioned Latent World Model for In-Cabin Dynamics Rollout 7. Urban Boundaries, Social Barriers: A Benchmark and Vision-Centric Framework for Mapping Gated Communities and Equity Implications 8. Less Tokens, Better Forecasts: Sparse Residual Routing for Efficient Weather Prediction 9. RainODE: Continuous-Time Precipitation Forecasting with Latent Neural ODEs 10. Pushing the Limits of High-Resolution Weather Forecasting through Data Scaling 11. PhenoLeaf-TS: A Time-Series Benchmark for Leaf Instance Segmentation, Tracking, and Growth Stage Classification |
|---|

此公众号的文章皆系本人原创,辛苦码字不易!如需转载,引用请注明出处。如需商用,联系作者。
点击文末阅读原文,跳转到笔者的知乎链接(跳转到论文链接更方便)。
1 Long-term Traffic Simulation via Structured Autoregressive Modeling
链接:https://eccv.ecva.net/virtual/2026/poster/5085
arXiv:http://arxiv.org/abs/2606.31209v1
代码:https://github.com/Sephirex-X/RosettaSim
作者:Lingyu XIAO ⋅ Zexin Feng ⋅ Xintao Yan
关键词:交通模拟,自动驾驶,LLM

2 ECoSim: Data Efficient Fine-Tuning for Controllable Traffic Simulation
链接:https://eccv.ecva.net/virtual/2026/poster/3709
arXiv:http://arxiv.org/abs/2607.00545v1
代码:https://github.com/bob020416/ecosim
作者:Yu-Hsiang Chen ⋅ WEI-JER Chang ⋅ Yi-Ting Chen ⋅ Masayoshi TOMIZUKA
关键词:可控交通模拟,高效微调

3 RoadBench: Benchmarking MLLMs on Fine-Grained Spatial Understanding and Reasoning under Urban Road Scenarios
链接:https://eccv.ecva.net/virtual/2026/poster/5365
arXiv:http://arxiv.org/abs/2511.18011v2
代码:https://github.com/tsinghua-fib-lab/RoadBench
作者:Jun Zhang ⋅ Xin Zhang ⋅ Jie Feng ⋅ Long Chen ⋅ Junhui Wang ⋅ Zhicheng Liu ⋅ Depeng Jin ⋅ Yong Li
关键词:空间推理,多模态大模型,benchmark

4 OpenEarthAgent: A Unified Framework for Tool-Augmented Geospatial Agents
链接:https://eccv.ecva.net/virtual/2026/poster/4838
arXiv:http://arxiv.org/abs/2602.17665v4
代码:https://github.com/mbzuai-oryx/OpenEarthAgent
作者:Akashah Shabbir ⋅ Muhammad Umer Sheikh ⋅ Muhammad Akhtar Munir ⋅ Hiyam Debary ⋅ Mustansar Fiaz ⋅ Muhammad Zaigham Zaheer ⋅ Paolo Fraccaro ⋅ Fahad Shahbaz Khan ⋅ Muhammad Haris Khan ⋅ Xiao Xiang Zhu ⋅ Salman Khan
关键词:地球观测,Agents

5 UrbanAlign: Post-hoc Semantic Calibration for VLM-Human Preference Alignment
链接:https://eccv.ecva.net/virtual/2026/poster/4316
arXiv:http://arxiv.org/abs/2602.19442v4
作者:Yecheng Zhang ⋅ RONG ZHAO ⋅ Zhizhou Sha ⋅ Yong Li ⋅ Lei Wang ⋅ Ce Hou ⋅ Wen Ji ⋅ HUANG HAO ⋅ yunshan wan ⋅ Jian Yu ⋅ Junhao Xia ⋅ Yuru Zhang ⋅ Chunlei Shi
关键词:人类偏好一致性,事后概念瓶颈,视觉语言模型,城市感知

6 Driver-WM: A Driver-Centric Traffic-Conditioned Latent World Model for In-Cabin Dynamics Rollout
链接:https://eccv.ecva.net/virtual/2026/poster/3775
arXiv:http://arxiv.org/abs/2605.05092v2
代码:https://fisher75.github.io/haozhuangchi.github.io/driver-wm/
作者:Haozhuang Chi ⋅ Daosheng Qiu ⋅ Hao Su ⋅ Haochen Liu ⋅ Zirui Li ⋅ Haoruo Zhang ⋅ Chen Lv
关键词:世界模型,驾驶员监控系统,VLM,人机交互驾驶自动化,自动驾驶

7 Urban Boundaries, Social Barriers: A Benchmark and Vision-Centric Framework for Mapping Gated Communities and Equity Implications
链接:https://eccv.ecva.net/virtual/2026/poster/5544
代码:https://github.com/MinweiZhao/GBA-GCs/
作者:Minwei Zhao ⋅ WEIMING ZHANG ⋅ Jiawang DU ⋅ Qiming LIU ⋅ Weiming Zhuang ⋅ Pei Nie ⋅ Cai Wu
关键词:公平性,兴趣区域(AOI)


8 Less Tokens, Better Forecasts: Sparse Residual Routing for Efficient Weather Prediction
链接:https://eccv.ecva.net/virtual/2026/poster/3283
arXiv:http://arxiv.org/abs/2607.02829v1
代码:https://github.com/janet-sw/Sparse-Reslim
作者:Janet Wang ⋅ Yunbei Zhang ⋅ Lin Zhao ⋅ Xi Xiao ⋅ Jihun Hamm ⋅ Xiao Wang
关键词:天气预报,vision transformer,高效训练

9 RainODE: Continuous-Time Precipitation Forecasting with Latent Neural ODEs
链接:https://eccv.ecva.net/virtual/2026/poster/4156
arXiv:http://arxiv.org/abs/2606.29855v1
代码:https://github.com/SeongYE/RainODE
作者:Yeeun Seong ⋅ Doyi Kim ⋅ Minseok Seo ⋅ Changick Kim
关键词:天气预报,持续时间建模,ODE

10 Pushing the Limits of High-Resolution Weather Forecasting through Data Scaling
链接:https://eccv.ecva.net/virtual/2026/poster/4737
arXiv:http://arxiv.org/abs/2608.14652v1
作者:Yang Zhao ⋅ Peisong Niu ⋅ Tian Zhou ⋅ Ziqing Ma ⋅ Guanlong Ma ⋅ Rong Jin ⋅ Huiling Yuan ⋅ Liang Sun
关键词:天气预报,scaling law,高分辨率

11 PhenoLeaf-TS: A Time-Series Benchmark for Leaf Instance Segmentation, Tracking, and Growth Stage Classification
链接:https://eccv.ecva.net/virtual/2026/poster/5291
代码:https://github.com/pisyntor/PhenoLeaf-TS
作者:Rijad Saric ⋅ Basim Azam ⋅ Sarmad Khan ⋅ Edhem Custovic
关键词:RGB时间序列,叶片生长
