第46卷第1期2023年2月电子器件ChineseJournalofElectronDevicesVol.46No.1Feb．2023项目来源:国家自然科学基金资助项目(61671141，61501038)收稿日期:2021－12－20修改日期:2022－02－14ＲesearchonIntegratedCircuitWeldingQualityMonitoringSystemBasedonImageProcessingTechnology*ZHANGLili，BAOXirong*，WANGTong，GUOJialin，YANGJinzhu(NationalComputerExperimentalTeachingDemonstrationCenter，SchoolofComputerScienceandEngineering，NortheasternUniversity，ShenyangLiaoning110819，China)Abstract:Withthedevelopmentofcomputertechnology，theapplicationofdigitalimageprocessingprovidesnewconditionsforthein-spectionofcircuitboardweldingquality．Theimageregistrationalgorithmofreferenceimageandimagetobedetectedisemphaticallystudied，Theimageregistrationmethodbasedongraylevelinformationisadopted．PVinterpolationmethodisusedtodirectupdatethejointhistogramandcalculatethemutualinformationvalueofreferenceimageandfloatingimage．ThenimprovedPowellalgorithmisusedtosearchtheoptimalregistrationparameters，andtheregistrationimageisobtainedthroughinterpolatingaccordingtotheregistra-tionparameters．Intheaspectofdefectdetection，colorrecognitionisusedtofindthedifferencepartsbetweentheregisteredimageandthereferenceimage，andhistogrammatchingandedgedetectionareusedtofilteroutthefalsedefectsandnoisepartstogetthedefectpart．Finallythedefectpositioniscalibrated．Experimentalresultsshowthattheproposedmethodcanachievefastandhigh-precisiondetectionofweldingdefects，thusfurtherreducinglaborcostsandimprovingdetectionefficiency．Keywords:imageprocessing;imageregistration;powellalgorithm;defectdetection;weldingqualityEEACC:6135doi:10．3969/j．issn．1005－9490．2023．01．010基于图像处理技术的集成电路焊接质量监测系统研究*张立立，鲍喜荣*，王彤，郭嘉林，杨金柱(东北大学计算机科学与工程学院，国家级计算机实验教学示范中心，辽宁沈阳110819)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...
