基于自适应神经模糊推理系统的纸病二次辨识

The Secondary Identification of Paper Defects Based on Adaptive Neural-fuzzy Inference System

关键词:  机器视觉  FPGA  纸病二次辨识  自适应神经模糊推理系统（ANFIS）

Key Words:machine vision  FPGA  secondary identification of paper disease  adaptive neural fuzzy inference system (ANFIS)

摘要:针对造纸生产线上宽幅高速纸机纸病检测系统快速性和精确性的挑战，提出“工业相机+FPGA（Field-Programmable Gate Array）+计算机”模式下的基于自适应神经模糊推理系统的纸病二次辨识方法。使用CCD相机采集纸张图像，通过FPGA完成图像预处理和一次辨识（粗辨识+过辨识）；计算机通过自适应神经模糊推理系统（ANFIS）对疑似纸病区域二次辨识（精确辨识），判断出纸病和种类。实验表明，该方法能够准确地辨识各种纸病。

Abstract:In order to ensure the contradiction between the speed and accuracy in paper defects detection on the wide and high speed paper machine, a method of paper defects secondary identification was put forward based on adaptive neural-fuzzy inference system (ANFIS) on the mode of “FPGA+computer”. The paper images collected by CCD cameras were completed image preprocessing and the first identification through FPGA; by using the ANFIS, the computer was able to complete the second identification of the suspected paper defects area to determine whether the paper defects existed and its type. Experiments showed that the method could accurately identify various paper defects.