| Vision inspection, as a promoting technology, has been applied widely in industrical area. This dissertation focuses on the inspection problems in Trim/Formming process in semiconductor BOL(Back of line) assembly. To solve these problems, an IC appearance inspection system is designed and developed by machine vision with digital image theory and method. An extensive study on system design, architecture and implement in crucial technologies is carried out. The main research as follows:1. Aimed at the problems of the ICs recognize and locate, a match strategy, sequential decision algorithm, is raised. Inspection accuracy is improved dramatically as well as just a little more time is needed. Applications show that the target image is effectively recognized and located and the proposed method is stable and successful.2. When coming to the IC leads inspection, LOG mask is convoluted with the polyline method to detect the edge. According to lead edge information, the dimensions of leads, such as lead count, pitch width, are calculated in the image coordinate with edge gray level change to judge the lead quality. This solution of inspection is proved to be practicable and feasible.The IC Leadframe Inspection System can detect most kinds of defects in the surface of chips. The actual requirement of the Enterprise has been satisfied by the system and it is also the application and validation of the theory of machine vision and Image processing.The critical analysis and application indicate that the research work done in this thesis is of practical value. |
PDF Full Text Request