布面疵点光学/计算机混合识别法的研究

Using Optical/Digital Hybrid Processor to Inspect Fabric Defects

  • 摘要: 验布是纺织工业中一道重要工序,但至今未能实观验布自动化,主要原因是布的疵点种类繁多不易识别。本文应用光学变换系统对疵点特征进行变换,用楔/环探测器进行特征提取和信息量压缩,应用计算机模式识别理论进行疵点的分类,经实验证明应用这种方法识别疵点能取得良好效果。

     

    Abstract: Fabric inspection plays a significant part in the textile industry, but it can not be done automatically yet, because the full number of defect fault is large and the defect cannot be inspected easily.In this paper, the defect fault information has been transformed by an optical system, sampled by a wedge/ring detector and then classified by a computer by using the pattern recognition theory. We have obtained a good result in our experiment.

     

/

返回文章
返回