基于图像处理的棒材识别计数系统设计与实现

发布时间：2018-03-30 01:07

本文选题：棒材图像　切入点：自动计数　出处：《郑州大学》2017年硕士论文

【摘要】：在棒材的交易过程中快速并准确的计算出棒材的数量尤其重要。目前市面上比较流行的计数方法主要有人工计数、射频标签、光电传感器。人工计数法不仅耗费人力,而且计数准确率不能达到百分百;射频标签法虽然计数准确率可达到百分之百,但是射频标签的耗费太大;光电传感器对于扎捆棒材来说,计数效果十分不理想;近几年图像领域的学者们研究出了一种棒材图像计数法,相对而言,图像计数法的成本少,准确率高,是目前棒材识别领域研究较多的一种方法。为了解决在棒材验收和出入库时快速有效的计数问题,本文设计了一种棒材自动识别计数系统,以凌动工控板为主板,配合USB摄像头作为系统的硬件平台,在Windows 7操作系统中以Visual studio 2013作为软件平台。针对棒材采集中常常面临到棒材部分遮挡、棒材间粘连等问题,采用一系列的预处理改善棒材图像的质量。通过对比采用传统的连通区域标记计数算法进行计数,并加入连通域面积对比、可识别标记的优化处理算法作为本文系统的核心算法,该算法可以完善传统连通域计数方法漏记、误判的现象。在凌动工控板上安装USB驱动程序,直接调用OpenCV库中的cap函数打开摄像头,通过键盘上的ESC按键进行图像采集和保存的控制端,在COME1054-IOBoard核心板上进行算法处理,调用计算机视觉库OpenCV3.0实现棒材图像的识别计数系统。本文实现以COME1054-IOBoard核心板为硬件平台,结合改进型的连通域计数法于一体的棒材图像计数系统。系统处理过的棒材图像中会明确的显示出识别出来及未识别出来的棒材,操作者可以通过加入人工干预的方法调整识别数量,使得识别计数的正确率达到百分之百。
[Abstract]:It is particularly important to calculate the number of bars quickly and accurately in the course of bar trading. At present, the most popular counting methods in the market are manual counting, radio frequency tags, photoelectric sensors. And the counting accuracy can not reach 100%, although the radio frequency tag counting accuracy can reach 100%, but the cost of radio frequency tag is too large, the counting effect of photoelectric sensor is not ideal for binding bar; In recent years, researchers in the field of image have developed a bar image counting method. Comparatively speaking, the image counting method has the advantages of low cost and high accuracy. In order to solve the problem of quick and effective counting in bar acceptance and storage, an automatic recognition and counting system for bar is designed in this paper, in which the ice industrial control board is used as the main board, in order to solve the problem of rapid and effective counting in the field of bar identification. With the USB camera as the hardware platform of the system, Visual studio 2013 is used as the software platform in the Windows 7 operating system. A series of preprocessing is used to improve the quality of bar image. As the core algorithm of this system, the algorithm can perfect the missing and misjudgment phenomenon of traditional connected domain counting method. The USB driver is installed on the ice industrial control board. The cap function in the OpenCV library is called directly to open the camera, and the control end of image acquisition and storage is carried out through the ESC key on the keyboard, and the algorithm is processed on the COME1054-IOBoard core board. The recognition and counting system of bar image is realized by calling computer vision library OpenCV3.0. The hardware platform of COME1054-IOBoard core board is realized in this paper. In combination with the improved connected domain counting method in one bar image counting system. The system processed bar image will clearly show the recognized and unrecognized bar, The operator can adjust the recognition quantity by adding manual intervention to make the correct rate of recognition count reach 100%.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH724;TP391.41

【相似文献】