基于多模式分类算法的棉花异性纤维自动检测系统的设计与实现

发布时间：2018-04-22 17:40

本文选题：异性纤维 + 自动检测系统　；参考：《河南科技大学》2017年硕士论文

【摘要】：近年来,由于全国棉花市场管理混乱,所收购的棉花中异性纤维严重超标,极大地影响了棉纺织产品的质量,从而也影响了我国棉纺行业的国内外的信誉,使棉纺织出口受到了很大的阻力。国外清除异纤的设备价格昂贵,且检出效果并不理想,国内有些企业虽已推出产品,但效果不好。为此,研制棉花异性纤维自动检测系统来改进棉花初步加工工艺,提高棉纺织品的成品质量,降低产品成本具有非常实际、重要的意义,同时也具有很好的市场价值和发展前景。本文在介绍了异性纤维分拣系统研究的背景及发展状况下,提出了以计算机图像处理技术为基础的异性纤维分拣算法,并对该算法的实现进行了比较深入的分析和研究,通过严格论证与分析,建立了图像处理的数学模型,即棉花在RGB空间的三维颜色数学模型,利用特征分析与类型归纳法,提出了针对普通异性纤维非线性双阈值算法和针对细小杂质的基于微分的图像处理算法相结合,提高了图形处理的效率,保证系统根据采集到的图像数据能够快速、准确的作出判断。实现了对白色丙纶丝和头发丝等细小杂质两方面检测的突破。而对于硬软件部分的实现则是合理搭建硬件系统平台,尤其是在相机的非线性问题的解决及光源的选择上,通过大量的实验以求得最佳的信号输入,使由硬件系统带来的不利因素达到了最小化。而软件系统则利用Visual C++6.0编制,在图像采集卡、CCD摄像机、C8051F340单片机等硬件基础上,充分发挥机器视觉技术的优势,实现了棉花异纤的在线、实时、高精度、非接触快速的动态检测及分拣。体现出了作为现代科学技术热点领域的机器视觉技术的诸多优点。最终经过多次现场调试,证明该系统具有检测速度快、分拣率高的特点。系统完全满足工业检测的要求。通过本文研究工作,研制并开发出具有自主知识产权的棉花异性纤维在线实时分拣系统,并使之成为应用于现场的工业级机器视觉在线检测设备。
[Abstract]:In recent years, due to the confusion in the management of the national cotton market, the heterosexual fibers in the cotton purchased have seriously exceeded the standard, which has greatly affected the quality of cotton textile products, thus affecting the domestic and foreign reputation of the cotton textile industry in our country. Cotton textile exports are subject to great resistance. The price of foreign equipment for removing foreign fibers is expensive, and the detection effect is not satisfactory. Some domestic enterprises have launched products, but the effect is not good. Therefore, it is of great practical and important significance to develop an automatic testing system for cotton heterosexual fibers to improve the preliminary processing process of cotton, to improve the quality of finished products of cotton textiles, and to reduce the cost of the products. At the same time also has a good market value and development prospects. In this paper, the background and development of heterosexual fiber sorting system are introduced, and an algorithm of heterosexual fiber sorting based on computer image processing technology is put forward, and the realization of this algorithm is analyzed and studied deeply. Through strict demonstration and analysis, the mathematical model of image processing, that is, the 3D color mathematical model of cotton in RGB space, is established, and the method of feature analysis and type induction is used. A nonlinear double threshold algorithm for ordinary anisotropic fibers and a differential based image processing algorithm for fine impurities are proposed to improve the efficiency of graphics processing and ensure that the system can be fast according to the collected image data. Make an accurate judgment. A breakthrough was achieved in the detection of fine impurities such as white polypropylene filament and hair filament. For the hardware and software part, the hardware system platform is set up reasonably, especially in solving the nonlinear problem of the camera and the choice of the light source, through a large number of experiments to obtain the best signal input. Minimize the disadvantages caused by the hardware system. The software system is programmed by Visual C 6.0, based on the hardware such as C8051F340 single chip computer and so on, and the advantages of machine vision technology are brought into full play. The on-line, real-time and high precision of cotton fiber are realized. Non-contact fast dynamic detection and sorting. It embodies many advantages of machine vision technology, which is a hot field of modern science and technology. Finally, after many field debugging, it is proved that the system has the characteristics of fast detection speed and high sorting rate. The system fully meets the requirements of industrial inspection. Through the research work in this paper, a cotton heterosexual fiber on-line real-time sorting system with independent intellectual property rights has been developed, and it has become an industrial machine vision on-line testing equipment applied in the field.
【学位授予单位】：河南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TS103.61;TP391.41

【参考文献】