摘要
随着计算机的发展,数字图像处理的应用为电路板焊接质量的检测提供了新的条件。重点研究了参考图像和待检测图像的图像配准算法。采用基于灰度信息的图像配准方法,使用PV插值法直接更新联合直方图并由此计算出参考图像和浮动图像的互信息值。然后采用改进的鲍威尔算法搜索最优配准参数,根据配准参数进行插值从而得到配准图像。在缺陷检测方面,通过颜色识别寻找配准图像和参考图像的差异部分,再通过直方图匹配和边缘检测滤去伪缺陷和噪声部分得到缺陷部分,最后标定缺陷位置。实验结果表明,采用所设计的方法可以实现焊接缺陷的快速并且高精度的检测,从而进一步降低人工成本,提高检测效率。
With the development of computer technology,the application of digital image processing provides new conditions for the in-spection of circuit board welding quality.The image registration algorithm of reference image and image to be detected is emphatically studied,The image registration method based on gray level information is adopted.PV interpolation method is used to direct update the joint histogram and calculate the mutual information value of reference image and floating image.Then improved Powell algorithm is used to search the optimal registration parameters,and the registration image is obtained through interpolating according to the registra-tion parameters.In the aspect of defect detection,color recognition is used to find the difference parts between the registered image and the reference image,and histogram matching and edge detection are used to filter out the false defects and noise parts to get the defect part.Finally the defect position is calibrated.Experimental results show that the proposed method can achieve fast and high-precision detection of welding defects,thus further reducing labor costs and improving detection efficiency.
作者
张立立
鲍喜荣
王彤
郭嘉林
杨金柱
ZHANG Lili;BAO Xirong;WANG Tong;GUO Jialin;YANG Jinzhu(National Computer Experimental Teaching Demonstration Center,School of Computer Science and Engineering,Northeastern University,Shenyang Liaoning 110819,China)
出处
《电子器件》
CAS
北大核心
2023年第1期57-61,共5页
Chinese Journal of Electron Devices
基金
国家自然科学基金资助项目(61671141,61501038)。
关键词
图像处理
图像配准
鲍威尔算法
缺陷检测
焊接质量
image processing
image registration
powell algorithm
defect detection
welding quality
