摘要
针对纸病检测的快速性与实时性问题,结合阈值法与形态学法的优势,设计了一种基于现场可编程门陈列(FPGA)的纸病检测系统。该系统采用了“CMOS相机+FPGA”的硬件模式,利用FPGA的并行流水线处理特性对数据流进行处理与识别,使图像预处理、特征提取、纸病识别与反馈并行完成,减少了额外的时钟消耗,且资源占用率极小。采用灰度阈值法与形态学法结合的方法提取特征值,有效地减少了光照对纸病检测的影响,提高了纸病的检测精度,且算法简单易于硬件实现。结果表明,该系统具有很强的实时性,能够快速准确地进行常见纸病的在线检测与辨识。
Aiming at the rapidity and real-time problem of paper disease detection,combined with the advantages of threshold method and morphological method,a paper disease detection system based on FPGA was designed.The system adopted the hardware mode of“CMOS camera+FPGA”,which used the parallel pipeline processing characteristics of FPGA to process and identify the data stream,which enabled image preprocessing,feature extraction,paper disease identification and feedback to be completed in parallel,reducing the consumption of extra clock,and had minimal resource occupancy.The feature value extraction used the combination of gray threshold method and morphological method,which could reduce the influence of illumination on paper disease detection effectively,and improved the detection accuracy of paper disease.The algorithm of this system was also simple and easy to implement in hardware.The test results showed that the system had strong real-time performance and could quickly and accurately perform on-line detection and identification of common paper diseases.
作者
赵睿轩
王思琦
周强
ZHAO Ruixuan;WANG Siqi;ZHOU Qiang(College of Electrical&Information Engineering,Shaanxi University of Science and Technology,Xi'an,Shaanxi Province,710021)
出处
《中国造纸》
CAS
北大核心
2019年第A01期103-108,共6页
China Pulp & Paper
基金
咸阳市科技计划项目(2017K02-06)。
