期刊文献+

位置信息约束的SMT料盘X射线图像检测方法 被引量:2

Method for detecting the X-Ray images of SMT materials plates based on the constraints of position information
在线阅读 下载PDF
导出
摘要 SMT封装物料盘的X射线透视图像发生局部粘连,导致分割准确性降低影响物料的检测与计数统计。提出了基于位置信息约束的分割检测方法,利用器件位置信息对分割区域进行约束提高分割与检测的准确率。首先依据元器件呈螺旋状排列规则,拟合物料盘中心点并计算最内环排列中起点器件位置信息;然后基于中心点与各元器件的位置约束模型,进行法向位置信息与先验位置信息双重约束,对目标器件所在区域进行限定;最后划分分割界限,完成对粘连目标的分割和对元器件的检测与统计。实验结果表明:该方法能够提高物料盘X射线透视图像粘连区域分割与检测的准确率,在有效像元素9 216 pixel,细节解析度110 1p/cm成像环境之下,对不同规格物料盘进行实测,检测误差率控制在0.15%之内。 The local adhesion in the X-Ray images of SMT material plates has negative effects on the accuracy of" counting and detection for components. A method based on the constraints of position information was proposed, which was used to improve the accuracy of segmentation and detection with the position information between components. Firstly, the center point and the starting components arranged in the inner ring were fit out on the basis of the components in a spiral arrangement rule. Then the dual constraints of normal position and priori position, which were based on the constraint model between the center point and different components, were finished to limit the region of target device. Finally, the segmentation for conglutination between components was completed by dividing the boundaries between components. Experimental results show that this method can improve the accuracy ofthe result of segmentation and detection. The detection error plates is controlled within 0.15% under the experimental element and detail resolution of 110 1p/cm. rate tested on different specifications material conditions of 9 216 pixels effective image
出处 《红外与激光工程》 EI CSCD 北大核心 2017年第6期160-166,共7页 Infrared and Laser Engineering
基金 国家自然科学基金(61405144) 天津市科技支撑计划重点项目(14ZCZDGX00033) 天津市科技特派员项目(15JCTPJC56300)
关键词 X射线透视图像 粘连 计数统计 分割检测 位置信息约束 X-Ray images adhesion counting statistics segmentation and detection constraints of position information
作者简介 耿磊(1982-),男,副教授,硕士生导师,主要从事测试测量技术及仪器方面的研究。Email:genglei@tjpu.edu.cn 通讯作者:肖志涛(1971-),男,教授,博士生导师,主要从事图像处理与模式识别方面的研究。Email:xiaozhitao@tjpu.edu.cn.
  • 相关文献

参考文献7

二级参考文献62

共引文献73

同被引文献20

引证文献2

二级引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部