Image Reverse Engineering Based On Computer Vision

In recent years, reverse engineering and its related techniques have been developing very quickly in realms such as machinery manufacturing, mould manufacturing, rapid prototyping, etc. Based on the object, its geometric model is re-construct and used for analysis and manufacturing. "Image reverse engineering based on computer vision" studied in this paper, which regards image as media and computer vision theories as guidance, obtains the digital model of the object and combines with numerical machining. The NC code produced can be used for manufacturing.The total design and the designs of each functional mould have been done according to the demand of system development. Commencing with image gathering, having been pre-processed by digital filter, image processing, image analysis and generation of NC code have been done. The NC code produced by the system can be stored as exterior document, which enhance the general use of the system.During the stage of image processing, "composite contour recognition technology" is applied and realized in this system, which makes the detection result continuous single pixel edge, and complete segmentation is fulfilled. "Rapid thresholding method", which can search the threshold automatically in the gray scale histogram, is designed during the process of image segmentation. Judgement by people avoided, automation degree and operation speed increased. During the stage of image analysis, contour tracking is used for the recognition and extraction of the edge element. Then it is saved by means of the data structure "chain code". The shape of the contour can be recovered wholly by the chain code. The cutter path file of NC machining is generated based on the chain code and can be applied to production process.Application program development has been done based on the Windows2000 operating system and in application of Visual C++ programming language. The result comes to the anticipant. "Image reverse engineering based on computer vision" is the combination of modern manufacturing technology and computer image technology. It involves much related knowledge and should be the inevitable result of mutual infiltration and crossed development of modern science and technology.