发表机构
Bangladesh University of Engineering and Technology（孟加拉工程技术大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出一种无需训练的双通粗到细流水线，结合Qwen3-VL-32B-Instruct、YOLO11x与BoT-SORT，在零样本约束下于ACCIDENT @ CVPR基准测试中实现22%相对优势，达成0.504的三方调和均值得分。
AI 中文摘要
交通监控摄像头持续捕捉事故，但将原始闭路电视（CCTV）视频转换为可明确事故发生时间、地点及碰撞类型的结构化事件记录，目前仍无法规模化实现。ACCIDENT @ CVPR基准在无标注真实训练数据的严格约束下评估该联合预测任务。本文提出一种无需训练的双通粗到细流水线，将冻结的Qwen3-VL-32B-Instruct视觉语言模型与YOLO11x目标检测、BoT-SORT跟踪相结合。第一通对完整视频片段进行稀疏采样，以锚定碰撞发生的时间点；第二通围绕该估计值的紧密窗口重新检查，使用标注了稳定车辆身份和归一化边界框坐标的帧，为模型提供同一场景的视觉叠加层和显式数值描述。在包含2027个片段的官方真实CCTV测试集上，本系统的三方调和均值得分为0.504，超过所有主办方发布的基线，包括最优多模型集成（0.412），相对优势达22%。
英文摘要
Traffic surveillance cameras capture accidents continuously, yet converting raw CCTV footage into structured event records that pinpoint when, where, and what type of collision occurred remains unsolved at scale. The ACCIDENT @ CVPR benchmark evaluates exactly this joint prediction under a strict constraint: no labeled real-world training data is available. We introduce a training-free, two-pass coarse-to-fine pipeline that pairs a frozen Qwen3-VL-32B-Instruct vision-language model with YOLO11x object detection and BoT-SORT tracking. A first pass sparsely samples the full clip to anchor the collision moment in time; a second pass re-examines a tight window around that estimate using frames annotated with stable vehicle identities and normalized bounding-box coordinates, which gives the model both a visual overlay and an explicit numeric description of the same scene. On the official 2,027-clip real-CCTV test set, our system achieves a three-way harmonic mean score of 0.504, surpassing all organizer-published baselines including the best multi-model ensemble (0.412) by a 22% relative margin.
CommentsAccepted at the AUTOPILOT Workshop, CVPR 2026, Denver, CO
