IRVAL
收藏资源简介:
IRVAL 是一个用于红外视频处理和时空视频超分辨率研究的高分辨率红外视频数据集。该数据集目前包含 8 个长红外视频,格式为 .avi。根据相关论文,IRVAL 包含 108,512 个视频帧,空间分辨率为 512×512。数据采用氧化钒(VOx)非制冷焦平面阵列探测器在长波红外(LWIR)波段采集。视频来自车载和固定监控平台,覆盖了城市街道、车辆、行人和路边建筑等真实场景。该数据集适用于红外视频处理、红外视频超分辨率、时空视频超分辨率和时间一致性建模等研究领域。数据集以原始视频形式发布,而非预生成的低分辨率/高分辨率对。使用时需遵守 CC-BY-4.0 许可协议。
IRVAL is a high-resolution infrared video dataset dedicated to research on infrared video processing and spatiotemporal video super-resolution. Currently, this dataset includes 8 long infrared videos in .avi format. According to relevant papers, IRVAL contains 108,512 video frames with a spatial resolution of 512×512. The data was collected using a vanadium oxide (VOx) uncooled focal plane array detector in the long-wave infrared (LWIR) band. The videos originate from vehicle-mounted and fixed surveillance platforms, covering real-world scenarios such as urban streets, vehicles, pedestrians, and roadside buildings. This dataset is applicable to research fields including infrared video processing, infrared video super-resolution, spatiotemporal video super-resolution, and temporal consistency modeling. The dataset is released as raw videos, rather than pre-generated low-resolution/high-resolution pairs. Users must comply with the CC-BY-4.0 license when using it.




