Rapid Mark Recognition Based On Edge Detection Technology In The Research And Implementation

AR (Augmented Reality) is a technology which can blend virtual object into the real world in a3D manner and enable users to be deeply immersed in it and be convinced the virtual object is an actual part of the real world. AR technology is widely applied in various fields such as national defense, medicine, education and entertainment industry, etc. As the applications of AR in different fields are expanded, higher requirements are raised with regard to the performance of the AR systems.AR technology based on artificial marks is the one being most popularly used in different cases. It detects and identifies the markers placed in the real world, then it registers the virtual object into the real-word based on the information it acquires through marker detection. Therefore marker detection plays a very important role in the whole AR system and it basically determines the performance of the AR system. Marker detection consists of image process and image analysis.The main purpose of this paper is to research the maker detection and marker identification, and to come up with a better marker system based on the previous research. The main work conducted in this paper is listed as follows:Firstly, a new artificial marker based on edge detection is proposed. Maker detection is the base of marker tracking and registering, because it provides the basic information needed. After analyzing the methods used by other popular markers and their performances, the article raises a new marker system which proved to be efficient and quick in detection and identification.Secondly, the article researches the tracking and registering technologies. After the maker is detected and identified, the transformation between the real world and the virtual world need to be calculated based on the information acquired in the process of identification, and this is tracking and registering. The article discusses the theories being applied in the tracking and registering and implements them in a new AR system. The experiments conducted in the article prove they are efficient.Thirdly, the article implements a new AR system which uses the new marker proposed previously. The new AR system consists of five modules including video inputting module, marker detection module, register module, picture rendering module and video outputting module. Experiments over the new system prove the new marker is up to the expectation of its design.