基于机器视觉的纸张叠层间接计数方法研究与应用
本文选题:机器视觉 + 纸张叠层计数 ; 参考:《湖南工业大学》2017年硕士论文
【摘要】:在包装、印刷部门业务调度过程纸张计量、分解等工序,目前采用的方法主要有人工测量和机械设备测量,由于人工计数受主观因素影响较大,机械设备易对纸张造成损坏,本文设计了基于机器视觉的纸张叠层间接计数系统,采用基于机器视觉辨识纸张及测量厚度,利用数显高度尺测量总高度计算纸张张数。本文主要的研究内容及相关结论包括:针对标准纸张叠层数量多,检测高度高,单视角成像难以获取整体信息,多视角全方位获取图像成本较高,分段采集图像后拼接式的计数方法,算法较复杂等问题,设计了采样式间接计数方法,首先用数显高度尺测量整垛纸的总高度,然后对叠层侧面上、中、下三个位置进行取样建立纸张压力厚度模型,通过视觉的方法计算一个标准张所占的厚度,最后根据总高度与单张厚度的比值,估算整垛纸张数量。实验证明该方法简单快捷,准确率高。针对用全局特征做图像分割后进行纸张计数存在阈值难以选取的问题,根据纸张边缘灰度跳变特征,设计算法检测标记之间的纸张数量,首先从标记的颜色分布特征分析,可以用灰度投影的方法判断计数的起止点,其次为了保护纸张边缘特性,选择保护边缘的平滑滤波器进行滤波处理,并用二维Gabor滤波器对纸张边缘进行增强,最后由于叠层纸张不可避免的存在突起或浮起、纸张松散、间隙宽大等问题,对纸张排列空间位置关系和纸张平均宽度进行分析和估算,有效控制了漏检、误检的情况,提高了计数准确率,计数准确率可以达到98%以上。最后设计了自动化的纸张叠层计数系统,并对不同批次的一系列相似叠层进行计数。测试结果表明对于检测高度在1米以上叠层纸张计数效果理想,计数准确率在99%以上,同时分析了产生误差的因素及优化方向。
[Abstract]:In the packaging, printing department business scheduling process paper measurement, decomposition and other processes, the main methods used at present are manual measurement and mechanical equipment measurement. Because manual counting is greatly affected by subjective factors, machinery and equipment are easy to cause damage to paper. An indirect paper counting system based on machine vision is designed in this paper. The paper is based on machine vision to identify the paper and measure the thickness. The paper tensor is calculated by using the digital display height ruler to measure the total height. The main research contents and related conclusions of this paper include: aiming at the large number of standard paper laminated, high detection height, single view imaging is difficult to obtain the whole information, and multi-view omnidirectional image acquisition cost is high. In this paper, an indirect counting method of collecting pattern is designed. Firstly, the total height of the whole stack paper is measured with digital display height ruler, and then, on the side of the stack, in the middle of the stack, The paper pressure thickness model was established by sampling in the next three positions, and the thickness of a standard sheet was calculated by visual method. Finally, according to the ratio of total height to single sheet thickness, the amount of paper stack was estimated. Experiments show that this method is simple and fast, and has high accuracy. In order to solve the problem that it is difficult to select the threshold of paper counting after image segmentation with global features, according to the grayscale jump feature of the edge of paper, an algorithm is designed to detect the number of papers between the labels. Firstly, the color distribution features of the labels are analyzed. The method of gray projection can be used to determine the starting and ending points of the count. Secondly, in order to protect the edge characteristics of the paper, the smooth filter with edge protection is selected for filter processing, and the edge of the paper is enhanced by two-dimensional Gabor filter. Finally, due to the inevitable problems such as bulge or floating, loose paper and wide gap, the paper arrangement spatial position relationship and the average width of paper are analyzed and estimated, which effectively controls the situation of missed detection and false detection. The accuracy of counting is improved, and the accuracy of counting can reach more than 98%. Finally, an automatic paper stack counting system is designed, and a series of similar layers of different batches are counted. The test results show that the counting effect of laminated paper is ideal and the counting accuracy is more than 99%. At the same time, the factors causing errors and the optimization direction are analyzed.
【学位授予单位】:湖南工业大学
【学位级别】:硕士
【学位授予年份】:2017
【分类号】:TS801.8;TP391.41
【参考文献】
相关期刊论文 前10条
1 武禾栩;钟俊;;嵌入式通信管理机配置软件设计[J];数字技术与应用;2015年09期
2 陈飞;朱海斌;贾康成;肖昌炎;;基于相机阵列的叠层纸张数量检测[J];计算机工程与应用;2015年13期
3 方超;谭伟;杜建洪;;基于纹理特征的太阳能晶片计数研究[J];信息与电子工程;2011年02期
4 石绘;余文勇;;新型烟标计数装置的设计[J];计算机工程与应用;2010年26期
5 孙颖超;李振海;;基于数学形态学的圆形标志定位[J];地理空间信息;2010年01期
6 吴晴;吴振华;郑榜贵;;形态学在道路图像边界检测中的应用[J];控制工程;2010年01期
7 王富治;黄大贵;;基于峰谷形态的纸张纹路分割算法[J];电子测量与仪器学报;2009年06期
8 闫海霞;赵晓晖;;基于数学形态学的边缘检测方法[J];计算机应用研究;2008年11期
9 郑光;陈幼平;余文勇;艾武;;基于数学形态学的纸张计数算法的研究[J];微计算机信息;2007年21期
10 李毅,阮秋琦;应用纹理分析的纸张计数算法[J];中国图象图形学报;2004年09期
相关硕士学位论文 前8条
1 蒋进;基于图像识别的纸张计数系统研究[D];广东工业大学;2015年
2 高素文;基于亚像素精度的立体图像匹配算法研究[D];广东工业大学;2014年
3 朱海斌;烟标叠层数量机器视觉检测仪的研究与开发[D];湖南大学;2013年
4 张金峰;基于图像处理的棒材自动计数系统的研究[D];东北大学;2010年
5 吴宝臣;卡把机结构与纸币图像识别与计数系统的研究[D];东北大学;2010年
6 张树芳;棒材自动计数图像处理方法应用研究[D];东北大学;2009年
7 刘颖;纸币图像在线计数算法研究[D];华中科技大学;2008年
8 刘燕燕;纸币非接触成像计数系统算法研究[D];华中科技大学;2007年
,本文编号:2047673
本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2047673.html