Clear Weather 恶劣天气物体检测数据集

Clear Weather 是一个用于挑战恶劣天气条件下对物体进行检测的数据集。数据集包括现实世界驾驶场景中的 12000 个样本和雾室内受控天气条件下的 1500 个样本。该数据集包括雾、雪和雨等不同天气条件下总共 10 万个物体的准确的 2D 和 3D 边界框标记。该数据集为信号增强、域适应、对象检测或多模态传感器融合的广泛算法提供了一个试验场，重点学习传感器之间的鲁棒冗余。

Clear Weather is a dataset developed to challenge object detection under adverse weather conditions. It includes 12,000 samples from real-world driving scenarios and 1,500 samples from controlled indoor weather conditions with fog. The dataset provides accurate 2D and 3D bounding box annotations for a total of 100,000 objects across various weather conditions such as fog, snow and rain. This dataset acts as a testbed for a wide range of algorithms including signal augmentation, domain adaptation, object detection and multimodal sensor fusion, with a focus on learning robust redundancy between sensors.

搜集汇总

背景与挑战

背景概述

Clear Weather 是一个专注于恶劣天气条件下物体检测的数据集，包含现实世界驾驶场景和室内受控天气条件下的样本，总计约13500个样本和10万个物体的2D和3D边界框标记。该数据集覆盖雾、雪和雨等多种天气条件，旨在支持信号增强、域适应、对象检测及多模态传感器融合算法的研究，特别强调传感器间的鲁棒冗余学习。