Research On Machine Vision Based Package Inspection System

Based on the machine vision technology, this paper develops a modular product package inspection system. The application in a cigarette package line demonstrates its feasibility.The hardware of the developed system consists of three parts: image acquiring, image processing and inspecting result output. The image acquiring part uses a LED strip to provide bright and steady light source. The system adopts six progressive scanning CCD cameras to take pictures simultaneously, and uses the industrial PC as the image processing unit. A friendly human-machine interface is created to display all operation information and inspected results as well. In order to pick out bad packages in a very tight time, a high speed electromagnetic valve and an air cylinder are deliberately selected.The system takes VC++6.0, a tool of Microsoft, as the development panel, and HALCON, a professional image process software of Machine Vision Technology. Many image process algorithms are employed to tackle some frequently-appearing flaw of package, such as the Grey Level Inspecting, Template Matching, Edge Angle Inspecting, Edge Misplacement Inspecting, Edge Distance Inspecting, and Bar Code Identifying. Through the flexible combination of above algorithms, the system realizes the package surface quality inspection.The developed software mainly includes the inspection system setting, inspecting as well as images browse. The system setting includes system triggering, running and the environment setting. The inspecting part realizes the package image processing, result display and decision-making. The image browse part browses the flaw image information for tracing back package errors.The developed system in this paper has been used in a cigarette packing line. The success of its application demonstrates that the system design is feasible and the software is steady and reliable. The system provides an alternative tool for product package inspection.