摘要
针对国外用户提出的车库卷闸门的非接触式位移实时监测要求,通过TensorFlow平台与OpenCV库设计了基于IPC监控摄像头的位移检测算法。选取特定标签并采用核相关滤波追踪器对卷闸门上的标签进行捕捉追踪,通过图像预处理等多种手段对图像进行抗干扰处理进行标签捕捉,测量标签的实时位移,在不同尺度特征及光照条件下进行了重复性实验。该算法对标签捕捉精准,实时追踪效果良好,位移和速度测量效果稳定,误差控制在3%以内。
According to the non-contact real-time displacement monitoring requirements of garage rolling gate proposed by foreign users,a displacement detection algorithm based on IPC monitoring camera is designed through Tensorflow platform and OpenCV library.The specific selected label and the nuclear correlation filter tracker are used to capture and track the label on the rolling gate.The image is anti-interference processed by various means such as image preprocessing to capture the label,measuring the real-time displacement.The repeatability experiments are carried out under multi-scale features and light conditions.The experimental results show that the proposed algorithm has accuracyinlabel capture,real-time tracking,stable displacement and velocity measurement,and the error is controlled within 3%.
作者
穆莉莉
汪晨灿
储汇
宋陈
MU Lili;WANG Chencan;CHU Hui;SONG Chen(Schoolof Mechanical Engineering,Anhui University of Science and Technology,Huainan 232001,China)
出处
《洛阳理工学院学报（自然科学版）》
2022年第4期64-70,共7页
Journal of Luoyang Institute of Science and Technology：Natural Science Edition
基金
安徽省重点研发项目(202004A07020046).
关键词
视觉捕捉
位移监测
物体追踪
visual capture
displacement monitoring
object tracking
