Research On Feature Recognition Of Natural Pattern In Augmented Assembly System

The augmented assembly system has played an important role in the modern mechanical manufacturing industry, and is an important part of virtual assembly technology. The virtual assembly system to achieve the simulation of assembly process, and let the designer simulate and test the mechanical product to find out its faultiness and complete product assembly scheme before it manufactured. The augmented assembly system improve assembly system interactivity and operability by augmenting real environments based on virtual assembly system.This paper analysis in deficiencies of the augmented assembly system to let the system recognize natural pattern for improving the interactivity of the system by proposing a new feature recognition method based on natural pattern in real environment. This paper related to the virtual assembly system, feature classification, computer vision, and image processing etc. This paper researches on character of classification method and digital image based on the natural pattern to propose a new recognition method by combining feature classification, computer vision and natural pattern. The recognition method convert the texture features to the necessary condition of the screw thread, and use it to recognize the texture features. The recognition system consists of two parts, one is for image processing, and the other one for pattern recognition. The former processes the original image to satisfy the later for analyzing, and the later extract the information of feature to match it with the necessary condition. At the end of the paper, the recognition system tests and verifies the recognition method.