摘要
随着信息化时代的到来以及智能电子行业的高速发展,复杂环境下的运动目标检测得到了广泛应用。运动目标检测方法的检测效果和运算量是关键问题。该文综述了运动目标检测方法的研究进展,包括背景差分法、光流法、帧差法、深度学习以及它们的改进方法;对每种方法的原理、优缺点进行了提取与对比,并总结了各算法的研究进程;展望了未来的研究趋势,通过对运动目标检测方法的综述和展望以期为后续研究提供一定的参考。
With the advent of the information age and the rapid growth of smart electronics industry,motion target detection in complex environments has been widely utilized.Detection effectiveness and computational workload of motion target detection methods are keys to the issue.This review delves into the advancements in the field of motion target detection methodologies,including background subtraction,optical flow,frame differencing,and deep learning approaches,as well as their improvement techniques.The principles,merits,and limitations of each methodology are extracted and compared,with a synthesis of the research trajectory for various algorithms.Furthermore,prospective trends in future investigations are anticipated,providing a forward-looking perspective on the domain.The review and outlook of the motion target detection methods aim to provide some reference for the subsequent research.
作者
张颜月
代显智
ZHANG Yanyue;DAI Xianzhi(College of Electronic Information Engineering,China West Normal University,Nanchong 637009,China)
出处
《四川轻化工大学学报(自然科学版)》
2025年第2期54-62,共9页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)
基金
教育部“春晖计划”合作科研项目(2018自编22)
四川省教育厅科研基金项目(18ZA0467)。
关键词
运动目标检测
复杂环境
背景差分法
光流法
帧差法
深度学习
motion target detection
complex environment
background subtraction method
optical flow method
frame difference method
deep learning
