包装彩盒检测中定位核提取算法

发布时间：2018-01-04 05:34

本文关键词：包装彩盒检测中定位核提取算法　出处：《计算机工程与设计》2016年11期 　论文类型：期刊论文

【摘要】：为降低现有包装印刷品检测过程中感兴趣区域(简称定位核)提取方法的复杂度,提高检测效率,提出一种基于多特征的定位核自动提取方案。通过输入选中的矩形区域,输出参数对比度、占空比、XY方向置信区间的取值范围及匹配系数拟合曲线的对称性,综合判断提取定位核的好坏。提出自动求取中心点坐标误差法对该方案的有效性进行验证。仿真分析结果表明,该方案可在100ms内自动提取定位核,成功率达到99.78%以上,明显优于已有方案,所提验证方法简单可靠。
[Abstract]:In order to reduce the complexity of the existing extraction methods of region of interest (referred to as localization core) in the process of packaging print detection, improve the detection efficiency. This paper presents an automatic location kernel extraction scheme based on multi-feature. By inputting the selected rectangular region, the output parameter contrast and duty cycle are obtained. The range of XY direction confidence interval and the symmetry of matching coefficient fitting curve. The effectiveness of the scheme is verified by the automatic center point coordinate error method. The simulation results show that the scheme can extract the localization core automatically within 100ms. The success rate is more than 99.78%, which is obviously superior to the existing scheme. The proposed verification method is simple and reliable.
【作者单位】：江苏科技大学电子信息学院;
【分类号】：TP391.41;TB487
【正文快照】： 0引言包装彩盒检测的目的是判断产品的外包装是否合格,包括对包装表面可能的缺陷检测及条码、字符识别等[1]。目前最常用的检测方法为模板匹配法[2]。好的检测模板可以大大提高检出率,减少拒检情况的发生,从而提高产品的整体质量。建模过程耗时最多的是用于匹配定位的定位核(R

【相似文献】