单微处理器控制的集成化光电混合神经网络模式识别系统  

INTEGRATED OPTO-ELECTRONIC HYBRID NEURAL NETWORK FOR PATTERN RECONGNITON CONTROLLED BY SINGLE MICROPROCESSOR

作　　者：孙杰[1] 翟宏琛[1] 张延昕[1] 张铁群[1] 刘钥 唐怡亮[1] 

机构地区：[1]南开大学现代光学研究所,教育部光电信息技术科学开放实验室,天津300071

出　　处：《南开大学学报（自然科学版）》2000年第1期106-109,118,共5页Acta Scientiarum Naturalium Universitatis Nankaiensis

基　　金：国家 8 63高技术研究计划 !( 863- 30 7- 0 2 ( 1 0 ) )资助项目

摘　　要：本文首次报道了应用单微处理器和单 LCD结构的光电混合神经网络模式识别系统 ,该系统使用了单微处理器控制的多屏幕显示及适用于多层光电神经网络的软硬件接口方法 ,充分利用了高性能 LCD的分辨率 ,提高了网络光电接口的数据传送速度及整机的集成化程度 。For the first time, an integrated opto electronic hybrid neural network for pattern recognition controlled by a single microprocessor is reported. By using the multi screen display controlled by the single microprocessor and by optimizing the software and the hard ware for opto electronic multilayer neural network, a full scope to the high resolution of the LCD and to the high transmitting rate of data through the opto electronic interfaces have been completed. As a result, the compatibility with the computers of high performance and the speed of the pattern recognition have been distinctly improved.

关 键 词：光电神经网络 微处理器控制 模式识别 集成化 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]