The Design And Implementation Of On-line Strip Surface Defect Inspection System Based On Computer Vision

In the cold-rolled strip production process, the surface quality control has been a major problem that troubles company much. There are many problems in traditional cold-rolled strip steel method, and the surface quality of steel can not be guaranteed. The use of computer vision technology for cold-rolled strip surface defects on-line detection emerges as a new type of integrated application technology in recent years. CAPL in BaoSteel is very sensitive for the holes especially on the edges which can result in breakage of the strip because of the thin strip and high tension. If such defects can be detected early and solved timely, the benefits are quite obvious.This thesis describes the design and implementation of inspection system for the strip surface quality defect, and the system is programmed with Visual C++ 6.0 and SQL Server 2000, implementing capture, process, storage and delivery of image data. Based on computer vision technology, by analizing real-time the data collected, the system can found defect on the strip surface and alarms. This system can reflect the shape, size and location of the hole while effectively detecting a hole. With the advanced technology adopted by the system, design requirements are met and system function is completed.First, the background information of this thesis is introduced and on-line testing technology based on computer vision is discussed. The system needs have been particularly analyzed, and the solution of hardware and software development is given. Then detailed design and implementation of the seven modules in this system are given, and flow charts are sketched. The image acquisition and testing module and image processing module are the key part of the system. At last the summary to this thesis and the outlook of the system are given for further study.