MWIRSTD
收藏资源简介:
MWIRSTD是由印度理工学院罗尔基分校、贝内特大学和视觉认知实验室私人有限公司合作创建的中波红外(MWIR)小目标检测数据集。该数据集包含14个视频序列,总计约1053张图像,标注了三种不同类别的小目标。数据集使用冷却的MWIR成像仪在真实环境中捕捉,旨在为研究人员提供一个真实的测试平台,以开发和评估小目标检测技术。MWIRSTD特别适用于国防应用,因为许多小目标在运动中表现出时空规律性,适合于目标检测和跟踪研究。
MWIRSTD is a mid-wave infrared (MWIR) small object detection dataset jointly developed by the Indian Institute of Technology Roorkee, Bennett University, and Visual Cognition Lab Private Limited. This dataset includes 14 video sequences, totaling approximately 1,053 images, with annotations for small objects belonging to three distinct categories. Captured using a cooled MWIR imager in real-world environments, this dataset aims to provide researchers with a realistic test benchmark for developing and evaluating small object detection technologies. MWIRSTD is particularly well-suited for defense applications, as numerous small targets exhibit spatiotemporal regularity during motion, making it ideal for object detection and tracking research.




