摘要
使用图像识别技术,提出了一种应用于环形工业生产线仪表数据采集与自动监控的系统架构,并结合实际的环形生产线平台,给出了详细的硬件设计和软件设计。针对工业现场仪表样本采集困难的弱点,为保障系统的识别率,利用虚拟样本改进的BP神经网络设计了模式分类器,并移植到FPGA上,发挥其并行计算能力,实现了对生产线的实时、高效监控。
A system to realize data acquisition and automatic monitoring on the circular production line is proposed based on image recognition. The de-signs of software and hardware are schemed in detail on the real circular production line. In order to guarantee the recognition rate of the system, the vir-tual samples is used to improve BP neural network for designing pattern classifier. Combined the improved BP neural network with the ability of FPGA＇ s parallel computing, the system can monitor the products＇ progress and quality.
出处
《电视技术》
北大核心
2014年第5期198-201,共4页
Video Engineering
