摘要
针对监控视频中动态目标获取受光照变化、摄像头抖动等因素干扰的问题,提出基于YOLOv3算法的监控视频动态目标自动识别方法。首先,对监控视频图像进行预处理,包括背景建模和差分计算,以优化图像背景;接着,提取视频中的特征;最后,结合YOLOv3算法对提取的特征进行检测和跟踪,实现动态目标的自动识别与实时检测。实验结果显示,所提方法显著提升了动态目标的检测精度和速度,证明该方法在应用上取得了更好的效果,适用于监控视频中的动态目标自动识别。
An automatic recognition method for dynamic targets in surveillance videos based on the YOLOv3 algorithm is proposed to address the interference issues caused by factors such as illumination changes and camera jitter during dynamic target acquisition in surveillance footage.Firstly,the surveillance video images are preprocessed,including background modeling and differential calculation,to optimize the image background.Subsequently,features are extracted from the video.Finally,the extracted features are detected and tracked using the YOLOv3 algorithm,enabling the automatic recognition and real-time detection of dynamic targets.Experimental results demonstrate that the proposed method significantly improves the detection accuracy and speed of dynamic targets,proving its superior application performance and suitability for automatic recognition of dynamic targets in surveillance videos.
作者
程荣森
CHENG Rongsen(Xinyang Agriculture and Forestry College,Xinyang,Henan 464000,China)
出处
《自动化应用》
2025年第14期21-23,共3页
Automation Application
关键词
侦查技术
目标检测
监控视频
YOLOv3算法
动态目标获取
光照变化
摄像头抖动
detection technology
target detection
survillance video
YOLOv3 algorithm
dynamic target acquisition
illumination changes
camera jitter
