摘要
为了解决钢板表面缺陷视觉检测系统中图像处理的瓶颈问题,采用基于FPGA的嵌入式处理系统完成大数据量、实时、在线的处理任务,从而满足高速、宽幅、高分辨率的检测要求。嵌入式系统以Altera公司最新的StratixFPGA作为核心处理器,在分析FPGA专有的硬件结构基础上,对图像处理中的模板卷积算法进行了优化设计,采用Verilog语言对算法完成建模与实现,并在QuartusII平台进行了仿真验证。实验与算法仿真证明嵌入式处理系统的可行性与工程实用性,从而表面缺陷检测系统中的图像信息实时处理这一关键问题得以解决。
In order to overcome the bottleneck of image processing in surface defect detection system for steel plate, embedded processing system is proposed to accomplish the task of bulk data real-time processing on line for satisfying high speed, broad width and high resolution inspection requirements. The latest Stratix FPGA of Altera Inc is chosen as kernel processor. After analyzing specific hardware architecture of FPGA, template convolution algorithm of image processing is optimally designed based on FPGA hardware, the modeling and realization is performed by taking use of verilog HDL, and the simulation and verification is achieved in Quartus Ⅱ software platform. After experiments and algorithms simulations, the feasibility and practicability of embedded system are verified. In conclusion, the key problem of signal processing is solved in surface defect detection system.
出处
《传感技术学报》
EI
CAS
CSCD
北大核心
2006年第3期694-696,701,共4页
Chinese Journal of Sensors and Actuators
基金
国家科技部科研院所专项基金资助(SY01-院所专项-01)
关键词
FPGA
嵌入式系统
实时处理
钢板
表面缺陷
FPGA
embedded processing system
processing on line
steel strip
surface defect
verilog
