基于图像处理的城市轨道交通异物检测系统研究

浦伦

Research on Urban Rail Transit Foreign Body Detection System Based on Image Processing

PU Lun
摘要:
提出了一种适用于城市轨道交通的异物检测系统,可实现15帧/s的图像采集和异物检测。根据相关性原理,对相邻多帧检测结果进行相关性计算,提高了算法的准确率和抗干扰能力。通过LoRa扩频通信,可实现远距离、高抗干扰的无线通信网络。该系统中,各个智能摄像头通过LoRa发送和接收与其相关的命令的同时,实现数据的转发,对远程摄像头的通信提供中继转发功能。该系统在节约人力成本的前提下,提高了城市轨道交通的安全性。
Abstracts:
A foreign body detection system for urban rail transit is proposed, which is capable of conducting image collection and foreign body detection at fifteen frames per second. The accuracy and robustness of the algorithm is improved with correlation calculation for detection results of consecutive frames. With LoRa, a spread spectrum communication protocol, steady long distance wireless communication is realized. In the system, each smart cam receives and sends data and commands with a server via LoRa. Meanwhile, these smart cams can relay messages for other cams to stabilize whole system communication quality. The system improves the urban rail transit safety under the premise of saving labor cost.
论文检索