Research On Image Recognition Method Of Foreign Fibers In Lint

The lint is the basic material of textile industry. Foreign fibers is non-cotton fibers or sundries which is mixed in the lint. Although less foreign fibers content in lint, the defects and the yarn breakage in spinning would seriously affect the quality of cloth. Therefore, researching a real-time and accurate online detection method has an important significance for improving the quality of lint.The image recognition method of foreign fibers in lint is researched in this paper. To improve the recognition rate and reduce the time consumption, the main contents are as follows:1. Completed the model selection of CCD camera, lens, lighting source, and realized the image acquisition with the simple found foreign fibers detection system.2. To identify the color foreign fibers from lint image, the three-dimensional RGB color space is transformed into two-dimensional space of hue and saturation under the HSI color space. An improved two-dimensional iterative Otsu algorithm is proposed based on analysis of Otsu algorithm properties.3. In this paper, a twice segmentation algorithm is applied to recognize the translucent foreign fibers in lint.4. The texture features is applied to the field of white foreign fibers identification in this paper, the two-dimensional Otsu segmentation method based on GLCM texture features image is used to the recognize white foreign fibers.5. To verify and improve algorithm conveniently, the interface of foreign fiber recognition system is designed.Experiments show that, the time consumption of improved two-dimensional Otsu algorithm is reduced to a tenth of the standard two-dimensional Otsu algorithm, which improves the real-time effectively. The segment method based on GLCM texture features image not only improves the recognition rate of white foreign fibers, but also reduces the time consumption by non-equidistant gray-scale compression.