LSOTB-TIR
收藏资源简介:
LSOTB-TIR是一个大规模、高多样性的热红外目标跟踪基准,由哈尔滨工业大学深圳研究院创建。该数据集包含1400个热红外序列,总计超过60万帧,并标注了超过73万个目标边界框。数据集分为评估集和训练集,旨在支持深度学习在热红外目标跟踪领域的应用,解决现有基准数据量小、对象种类少、场景和挑战有限的问题。LSOTB-TIR的应用领域广泛,包括视频监控、海上救援和夜间驾驶辅助等,特别是在完全黑暗的环境中跟踪目标。
LSOTB-TIR is a large-scale, high-diversity thermal infrared object tracking benchmark developed by the Shenzhen Research Institute of Harbin Institute of Technology. This dataset comprises 1400 thermal infrared sequences, with a total of over 600,000 frames and more than 730,000 annotated target bounding boxes. Split into an evaluation set and a training set, this benchmark is designed to facilitate the application of deep learning in the field of thermal infrared object tracking, and address the shortcomings of existing benchmarks including small dataset scale, limited object categories, and narrow coverage of scenarios and challenge types. LSOTB-TIR has broad application prospects across multiple fields, including video surveillance, maritime rescue, nighttime driving assistance and other scenarios, particularly for target tracking in fully dark environments.




