基于Otsu算法的输送带撕裂视觉检测系统研究

发布时间：2018-05-24 06:21

本文选题：输送带 + 撕裂检测　；参考：《太原理工大学》2017年硕士论文

【摘要】：由于带式输送机具有大运量、高速度、不间断运输的优点,其被广泛应用于散状、块状物料运输系统。在煤炭生产、运输过程中,带式输送机也是同样拥有其他运输设备无可比拟的优势。因此,带式输送机是重要的煤炭生产、运输设备之一。带式输送机长期、高负荷运行过程中,输送带面临着托辊、滚筒故障摩擦、金属物件卡阻刺穿、矸石划伤等造成的输送带撕裂危险。输送带撕裂往往造成煤炭运输系统瘫痪和货物洒落堵塞巷道事故。严重情况更会导致人员伤亡事故,对煤炭安全生产造成了不利影响。本课题在深入了解对机器视觉系统的基础上,研究多种图像处理算法在输送带撕裂视觉检测系统应用的可行性,旨在设计出一套实用性更强的输送带撕裂视觉检测系统,为煤炭生产、运输系统保驾护航。本文主要研究主要包括:1、目前,输送带撕裂视觉检测方法相比较于其他输送带撕裂检测方法具有无法比拟的优越性。结合现有机器视觉系统在输送带撕裂检测领域的应用研究,提出一种实用性、可靠性、稳定性、实时性更强的基于Otsu算法的输送带撕裂视觉检测的研究方案。2、在深入分析研究机器视觉理论、结构的基础上,根据煤炭生产检测系统的设计原则,提出基于Otsu算法的输送带撕裂视觉检测系统设计方案,并深入研究系统结构、功能以及实现系统功能所需全部硬件设备。3、图像处理算法研究是输送带撕裂视觉检测系统设计的核心。课题针对工业CCD相机采集的灰度图像进行以下三部分算法研究:1图像预处理算法;2图像处理算法;3撕裂诊断算法。系统采用中值滤波和直方图均衡化对输送带表面图像进行预处理,达到去除椒盐噪声、增强图像对比度的目的。图像阈值分割处理是系统对输送带表面图像处理的方法,是系统撕裂检测的关键步骤。课题通过研究局部阈值分割算法、迭代阈值分割算法、Otsu阈值分割算法,综合比较三者图像处理时间、图像处理效果方面的优劣,确定Otsu阈值分割算法作为输送带撕裂视觉检测系统的图像处理算法。最后根据输送带表面二值图像的特点规定撕裂诊断规则、算法,达到系统撕裂检测的目的。4、提出输送带撕裂视觉检测系统软件设计结构流程,分析如何通过软件实现系统功能。虚拟仪器技术不断发展,模块化硬件结构设计,简洁的图形化Lab VIEW编程环境,为机器视觉系统带来了一次全新的革命。基于上述优点,输送带撕裂视觉检测系统使用虚拟仪器实现系统图像处理、数据通信、数据库存储等功能。5、依托实验室中现有的带式输送机设备,针对系统实现功能选择合适的工业CCD相机、LED光源、速度传感器、下位机PXI平台、上位PC机等硬件设备搭建实验平台。在带式输送机5m/s的运行速度下,启动系统,通过系统界面检验运行状态下系统对输送带表面图像处理效果,分析、对比系统输出的正常、撕裂输送带表面二值图像数据,设定撕裂诊断算法参数,证明系统的可行性。本系统是机器视觉在输送带撕裂检测领域的应用研究,系统通过CCD相机获取输送带表面图像,并综合应用中值滤波、直方图均衡化、Otsu灰度阈值分割方法对其处理、撕裂诊断,实现了对运行状态下输送带实时无损检测的功能。
[Abstract]:As the belt conveyor has the advantages of large volume, high speed and uninterrupted transportation, it is widely used in bulk and bulk material transportation system. In the process of coal production and transportation, belt conveyor is also unparalleled with other transportation equipment. Therefore, belt conveyor is one of the important coal production and transportation equipment. In the long and high load running process of the conveyor, the conveyor belt is faced with the roller, the friction of the roller failure, the sticking of the metal objects, the damage of the gangue and so on. The tear of the conveyor belt often causes the coal transportation system to be paralyzed and the goods are sprinkled and blocked in the roadway. The whole production has caused a negative impact. On the basis of understanding the machine vision system, this paper studies the feasibility of the application of multiple image processing algorithms in the conveyer belt tearing visual detection system, aiming at designing a set of more practical belt tear visual detection system for coal production and transportation system. The main research is as follows: 1, at present, the visual detection method of conveyer belt tearing is superior to other conveyor belt tearing detection methods. Combined with the application of existing machine vision system in the field of conveyer belt tearing detection, a practical, reliable, stable and more real-time based Otsu algorithm is proposed. .2, based on the analysis and study of the theory of machine vision and the design principle of the coal production detection system, the design scheme of the belt tear visual detection system based on Otsu algorithm is proposed, and the system structure, function and all hardware functions required to realize the system function are deeply studied. .3, the research of image processing algorithm is the core of the design of conveyor belt tearing visual detection system. The following three algorithms are studied for the gray image collected by industrial CCD cameras: the 1 image preprocessing algorithm, the 2 image processing algorithm and the 3 tearing diagnosis algorithm. The system uses median filter and histogram equalization to get the image of the conveyer belt. Preprocessing to remove salt and pepper noise and enhance image contrast. Image threshold segmentation is a system for processing the surface of the conveyer belt. It is the key step of the system tearing detection. By studying the local threshold segmentation algorithm, the iterative threshold segmentation algorithm and the Otsu threshold segmentation algorithm, the three image parts are compared synthetically. According to the advantages and disadvantages of time and image processing effect, the Otsu threshold segmentation algorithm is used as the image processing algorithm for the transmission band tearing visual detection system. Finally, according to the characteristics of the two value image of the conveyer belt, the tearing diagnosis rules are stipulated and the algorithm is used to achieve the purpose of the system tearing detection. The software of the transmission band tearing visual detection system is set up. This paper analyzes how to implement the system function through software. The development of the virtual instrument technology, the design of modular hardware structure and the simple graphical Lab VIEW programming environment have brought a new revolution to the machine vision system. Based on the above advantages, the conveyer belt tearing visual detection system uses virtual instrument to realize the system image Processing, data communication, database storage and other functions.5, relying on the existing belt conveyor equipment in the laboratory, to select the appropriate industrial CCD camera, LED light source, speed sensor, PXI platform of the lower computer, the upper PC machine and other hardware equipment to set up the experimental flat. Over the system interface test operation state of the system on the conveyor belt surface image processing effect, analysis, comparison of the normal output of the system, tearing two value image data on the conveyor belt surface, setting up the tearing diagnosis algorithm parameters, proving the feasibility of the system. This system is the application of machine vision in the field of conveyor belt tearing detection, the system is through the CCD phase The machine obtains the image of the conveyor belt surface, and uses median filter, histogram equalization, Otsu gray threshold segmentation method to handle and tear diagnosis, and realizes the function of real-time nondestructive testing of the conveyer belt in the running state.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】