基于行扫描的炮弹引信铆点特征提取  

A Feature Extraction Algorithm of Cannon-ball Fuse Riveted Point Based on Line Scanning

作　　者：郝旺[1] 杨润泽[1] 尹晓春[1] 

机构地区：[1]军械工程学院基础部,河北石家庄050003

出　　处：《计算机测量与控制》2011年第3期613-616,共4页Computer Measurement &Control

摘　　要：在炮弹引信铆点机器视觉识别中,有效消除图像中各种干扰、正确提取铆点特征是关键,解决好铆点特征的准确提取问题,是炮弹引信安全破除铆点的前提条件;根据铆点的结构特点,分析了产生各种干扰的原因及特点,在高斯滤波、阀值分割、形态学运算等基本运算的基础上,提出了一种基于行扫描的铆点特征提取方法;测试结果表明,提出的方法可行,铆点正确识别率达到90%以上;具有高效性和很好的鲁棒性的铆点特征提取,使得炮弹铆点的准确破除成为可能。In the process of Cannon--ball Fuse Riveted Point recognizing by machine vision, to eliminate the image interference effectively and to extract the riveted point feature accurately is the key point. The solution of Riveted Point recognizing problem, is the condition of destroying Riveted Point safely. Researching the reason of interference and its characteristic, after using basic image processing method by linear Gauss filtering, threshold segmentation, mathematical morphology, a line scanning algorithm is presented. This algorithm can resolve the problem of riveted point recognizing successfully. The recognizing accuracy ratio can reach 90 percents. To extract the riveted point feature effectively and robustly, make the operation of Fuse Riveted Point can be realized.

关 键 词：铆点 高斯滤波 阀值分割 行扫描 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]