Slave Machine Design Of Networking Entrance Guard System Based On Intelligent Image Card

Networking Entrance Guard System (NEGS) always is a kind of product people focus on researching and developing. Presupposing no effect on safety and reliability, the system can supply efficient centralized management mechanism of multi-entrance & friendly service circumstances to users, and create essential condition for raising economic & social benefits.With Intelligent Image Card (IIC) as key, NEGS is a new-pattern centralized management architecture of multi-entrance. Following Distributed Control System (DCS) design pattern and taking"Entrance-Lock-Key"relationship as mathematic model, NEGS is composed of one Host Machine, several Slave Ones and RS232/485 Comunication Bus.It forces entrance guard working remotely by distributing IIC characters on Host Machine and recognizing it on Slave Machine.This paper especially focuses on discussing both design and implementation of hardware and software of Slave Machine.Its hardware is composed of following modules:main controlling circuit,image sampling subsystem,compatible multi-lock driving port,programmable and energy conservative power,networking interface,timing circuit and so on.And it achieve functions like IIC image sampling, lock driving, voltage transforming & electricity supplying, remote power on, continuous timing and so on. Its software is composed of following subsystems: monitoring, image sampling & processing, IIC recognizing, IIC characters managing, intelligent parameters implementing,peripheral driver and so on. And it achieve functions like image sampling,pattern matching,charcters array management,intelligent option choosing,peripheral driving and so on.Aiming at networking feature, this paper proposes an application implementing SVM mulit-class classification distributely in IIC recognition with NEGS's network feature. First, the application takes IIC image sampling characters as samples to train SVM multi-class classifier on host machine.Second, it sends descriptive classifier parameters generating in last step reliably to corresponding slave machine by network.Third, takes the same IIC image sampling characters acquiring on slave machine as fact values into classifer to recognize.Experimental results and analysis proved that this application not only reduce load of slave machine in IIC recognition, but also raise recognizing correct rate & speed,and raise NEGS's overall performance indirectly. Its fundmental also can be applied in else NEGSs which takes else patterns recognition results as key and enlarge applications of SVM multi-class classification.