【摘要】 小目标检测一直以来都是计算机视觉领域极具价值且有挑战性的任务。现有的大多数检测算法不能很好地解决小目标检测困难的问题。小目标检测技术广泛应用于卫星遥感、智慧交通、国防安全和工业自动化等领域,具有重要的实用价值。针对近年来基于深度学习的小目标检测研究成果,对小目标检测这一研究热点进行了系统全面的分析与总结,综述了利用数据增强、超分辨率检测、特征增强和改进损失函数等提高小目标检测性能的方法,最后总结了小目标检测未来的研究方向。更多还原

【Abstract】 Small object detection is a challenging and valuable task in the field of computer vision.Most of the existing detection algorithms can not solve the difficult problem of small object detection.Small object detection technology is widely used in satellite remote sensing,intelligent transportation,national defense security and industrial automation and other fields,and has important practical value.Given the research results of small object detection based on deep learning in recent years,a systematic and comprehensive analysis and summary of the literature is conducted from the four perspectives:data enhancement methods,super-resolution methods,feature enhancement methods,and special loss function methods.Finally,the future research direction of small object detection is summarized.更多还原