arXiv 2603.162612026-03-17cs.CVcs.AI

AW-MoE：面向所有天气条件的专家混合方法用于鲁棒多模态3D目标检测

AW-MoE: All-Weather Mixture of Experts for Robust Multi-Modal 3D Object Detection

Fujian Key Laboratory of Urban Intelligent Sensing and Computing（福建城市智能感知与计算重点实验室）；School of Informatics, Xiamen University（厦门大学信息学院）；Zhongguancun Academy（中关村学院）

Hongwei Lin, Xun Huang, Chenglu Wen, Cheng Wang

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出AW-MoE，通过引入天气感知路由和统一双模态增强方法，提升多模态3D目标检测在恶劣天气下的鲁棒性，实验表明其在恶劣天气下的性能提升达15%。

摘要：

在恶劣天气条件下实现鲁棒的3D目标检测对于自动驾驶至关重要。然而，现有方法通常简单地将所有天气样本合并训练，而忽视了不同天气场景之间的数据分布差异，导致性能冲突。为了解决这一问题，我们引入了AW-MoE框架，创新性地将专家混合（MoE）整合到天气鲁棒的多模态3D目标检测方法中。AW-MoE结合了图像引导的天气感知路由（IWR），利用图像特征在不同天气条件下的优越判别性和对场景变化的不变性，实现精确的天气分类。基于此准确分类，IWR选择处理数据差异的前K个天气特定专家（WSE），确保在所有天气条件下都能获得最佳检测效果。此外，我们提出了统一双模态增强（UDMA）方法，用于同步激光雷达和4D雷达双模态数据增强，同时保持场景的真实感。在真实世界数据集上的广泛实验表明，AW-MoE在恶劣天气下的性能相比现有最先进方法提高了约15%，且推理开销可忽略不计。此外，将AW-MoE集成到现有基线检测器中，能够获得超越当前最先进方法的性能提升。这些结果展示了我们AW-MoE的有效性和强大的可扩展性。我们将在https://github.com/windlinsherlock/AW-MoE上公开代码。

英文摘要：

Robust 3D object detection under adverse weather conditions is crucial for autonomous driving. However, most existing methods simply combine all weather samples for training while overlooking data distribution discrepancies across different weather scenarios, leading to performance conflicts. To address this issue, we introduce AW-MoE, the framework that innovatively integrates Mixture of Experts (MoE) into weather-robust multi-modal 3D object detection approaches. AW-MoE incorporates Image-guided Weather-aware Routing (IWR), which leverages the superior discriminability of image features across weather conditions and their invariance to scene variations for precise weather classification. Based on this accurate classification, IWR selects the top-K most relevant Weather-Specific Experts (WSE) that handle data discrepancies, ensuring optimal detection under all weather conditions. Additionally, we propose a Unified Dual-Modal Augmentation (UDMA) for synchronous LiDAR and 4D Radar dual-modal data augmentation while preserving the realism of scenes. Extensive experiments on the real-world dataset demonstrate that AW-MoE achieves ~ 15% improvement in adverse-weather performance over state-of-the-art methods, while incurring negligible inference overhead. Moreover, integrating AW-MoE into established baseline detectors yields performance improvements surpassing current state-of-the-art methods. These results show the effectiveness and strong scalability of our AW-MoE. We will release the code publicly at https://github.com/windlinsherlock/AW-MoE.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。