视觉-热像相机数据集

视觉-热像相机数据集由自主机器人实验室创建，旨在研究视觉和热像数据的跨模态关联，特别是在无校准信息的情况下实现像素级对齐。数据集包含四个子集，分别来自汽车和地下无人机探测任务，涵盖日间和夜间条件。创建过程中，采用了Mattes互信息度量进行图像对齐，无需内部或外部校准参数。该数据集适用于机器人导航和环境认知研究，特别是在视觉退化环境中的应用，旨在解决多模态数据融合的问题。

The Visual-Thermal Camera Dataset was developed by the Autonomous Robotics Lab to investigate cross-modal correlations between visual and thermal data, particularly for achieving pixel-level alignment without any calibration information. The dataset consists of four subsets collected from automotive and underground drone detection tasks, covering both daytime and nighttime scenarios. During the dataset construction, the Mattes Mutual Information metric was adopted for image alignment, eliminating the need for internal or external calibration parameters. This dataset is applicable to research on robotic navigation and environmental cognition, especially for applications in visually degraded environments, and aims to address the challenges of multimodal data fusion.