基于计算机视觉的笔杆笔夹对正系统设计  

Design of Alignment System of Pen Shaft and Clamp Based on Computer Vision

作　　者：陈永清[1] 陈廉清[1] 蒋惠芬 谢建灿 郭建亮[1] 程晓民[1] 

机构地区：[1]宁波工程学院材料学院,浙江宁波315016 [2]杭州雷恩液压设备制造有限公司,浙江杭州311100 [3]衢州杭甬变压器有限公司,浙江衢州324000

出　　处：《测控技术》2016年第5期53-56,65,共5页Measurement & Control Technology

基　　金：国家自然科学基金资助项目(51275251);浙江省自然科学基金资助项目(Y14E050021)

摘　　要：针对制笔行业传统人工对正笔杆笔夹效率低、易遗漏、准确度差等弊端,提出了计算机视觉自动对正检测技术。笔的种类繁多,笔杆LOGO图案复杂、色彩多样,图像特征不同,采用了先作多尺度多结构元素相结合的形态学增强后进行灰度模板匹配的方法。通过对不同笔杆图像处理表明,能够抑制噪声及可疑边缘,保留LOGO图案的有用边缘信息,具有极强的实用性。基于该方法开发的自动对正检测系统,经试运行,结果表明该系统的精确度及效率均满足生产要求,且对正结果不受主观因素的影响,稳定性好、效率高。To overcome the shortomings of the low efficiency,omission-prone,and poor accuracy of traditional manual approach of penholder alignment,an automatic alignment method is proposed based on computer vision checking technique.Due to the wide varieties of pen,the complexity of penholder LOGO pattern,the diversification of colors,and the difference of image characteristics,the morphology with combination of multi-scale and multiple structural elements is adopted followed by matching grayscale template.Processing of different penholder image shows that it is capable of damping noise and suspicious edges as well as retaining useful edge information,which results in a high practical utility.Results of trial running for automatic alignment checking system developed with the proposed approach show that this system can meet the requirements in terms of accuracy and efficiency.Furthermore,the alignment result has high stability and feasibility due to the fact that it is not influenced by subjective factors.

关 键 词：形态学 图像处理 对正 计算机视觉 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]