基于近红外成像技术的棉花异纤检测

Cotton Foreign Fiber Detection Based on Near-infrared Imaging Technology

探讨基于近红外成像技术的棉花异纤检测方法。在近红外波段对混有12种异纤样本的棉花进行扫描成像并用线阵CCD相机获取图像,采用OpenCV图像算法进行处理,实现异纤检测;并对检测结果和异纤灰度特性进行分析。结果表明:有色异纤样本在各波长下与背景有较为明显的灰度差异,其中只有头发在905 nm波长下未被检测到,而其他有色异纤都可以被检测出来;白色或透明异纤样本在808 nm波长下检测效果较好,检测出了7种异纤样本中的5种,未出现错误检测。认为:在近红外波段检测白色或透明异纤是可行的,具有较好的检测效果。

The cotton foreign fiber detection based on near-infrared imaging technology was discussed.Scan imaging was performed for the cotton mixed with twelve kinds of foreign fiber samples in near-infrared band and linear array CCD camera was used to obtain the image.OpenCV algorithm method was adopted for processing.The foreign fiber detection was achieved.The test results and the foreign fiber gray scale characteristics were analyzed.The test results showed that the color foreign fibers had comparatively obvious gray scale difference from the background in every wave length.Among them,only hair was not detected in 905 nm wave length while other foreign fibers all could be detected.For the white or transparent foreign fiber samples,the detection effect was better in the wave length of 808 nm.Five foreign fibers in seven foreign fiber samples were detected and no error detection showed up.It is considered that the detection of white or transparent foreign fibers in near-infrared band is feasible and the detection effect is better.