基于机器视觉的棉花异性纤维识别方法

Cotton Foreign Fibers Identification Based on Machine Vision

本文提出了一种基于机器视觉识别原棉异性纤维的新方法。首先采集含异性纤维的原棉彩色图像,将其分解为三通道图像。其次,采用二维伽马函数来解决三通道图像光照不均匀问题。在此基础上,采用大律法对三通道图像进行分割。最后,将分割后的三通道图像进行融合,能正确识别出棉花样本中的异性纤维。实验结果表明:本文提出的检测系统效率高,其检出率高于95%,能够满足棉花工业应用的要求。

This paper proposed the new method for cotton foreign fibers identification based on machine vision.First of all,the color images of cotton foreign fibers are acquired and divided three channels images.Then,the 2D Gamma function is adopted to correct non-uniform illumination for images.On this basis,the big law method is used to segment the image.Finally,the segmented three channel images are fused,and the foreign fibers in the cotton samples can be correctly identified.The experimental results show that the proposed detection system has high efficiency and its detection rate is more than 95%,which can meet the requirements of cotton industrial application.