摘要
在复杂环境下提高地铁闸机智能识别系统的识别率是一项极其困难的工作。针对地铁闸机通行中各种事件的不同情况,提出了一种基于支持向量机的智能识别方法,并设计了基于双CPU控制和红外传感器的智能检测识别系统,通过合理安排红外传感器的位置,获取通行乘客经过闸机通道时的运动序列,结合图像识别等辅助功能,最终判断通行乘客情况。大量样本数据表明,通过SVM的方法使控制系统精度和处理能力得到提高,识别率超过90%,较好地满足了市场需求。
It is an extremely difficult task to improve the recognition rate of the intelligent recognition system of metro gates in a complex environment.In line with the different situations of various events in the movement of metro gates,an intelligent recognition method based on support vector machines is proposed,and an intelligent detection and recognition system based on dual CPU control and infrared sensors is designed,and the position of infrared sensors is reasonably arranged to obtain passengers’gate-passing movement sequence,and finally the situation of passing passengers in combination with auxiliary functions such as image recognition is judged.A large number of sample data show that the accuracy and processing capacity of the control system are improved by the SVM method,and the recognition rate reaches more than 90%,which fairly satisfies the market demand.
作者
张智华
孔维灿
曹伟
康泽宇
徐广泽
ZHANG Zhi-hua;KONG Wei-can;CAO Wei;KANG Ze-yu;XU Guang-ze(School of Transportation Engineering,Jiangsu Shipping College,Nantong 226010,China)
出处
《南通航运职业技术学院学报》
2021年第1期32-36,共5页
Journal of Nantong Vocational & Technical Shipping College
基金
江苏省大学生创新计划训练项目(202012703010Y)。
关键词
闸机
目标识别技术
双CPU
支持向量机
metro gate
target recognition technology
dual CPU
support vector machine
