Research Of Reconstruction And Camera Pose Estimation On Mobile Phone

Augmented Reality (Augmented Reality referred to as AR) technology is a hot topic in recent research at home and abroad, is the reality of seamless integration with the virtual scene, the scene after fusion in which the user is difficult to distinguish between virtual and reality, and ultimately enhance the user of the real scene user experience. With computer technology, and smart phones, PDA, tablet PCs and other intelligent terminal device development, it is increasingly dependent on these portable devices and mobile platforms Smart Augmented Reality applications will be more and more rich. 3D registration (also called the actual situation registration) has been enhanced is the core technology. With the rapid development in recent years, augmented reality, mobile platforms and the reality on the registration application of augmented reality has become a research hotspot.In this paper, most of the mobile augmented reality system on a single scene, and dependent on the marked block of the present situation of a faster, more robust and does not rely on said block of augmented reality system. The system greatly expands the scope of application of augmented reality technology, while addressing the cell phone because the memory is small, the weak can not achieve real-time computing power to run the problems and improve the accuracy of 3D registration and robustness.Implications of these issues, on the phone to achieve some of the traditional algorithms, and proposed some new algorithms and solutions, the main contents include:1) We proposed a 3D scene based on multi-up technology, not only to complete the case without marking up the scene of 3D, and solved a big scene in the scene and the tracking problem of mobile phone memory is limited.2) We proposed the classification of clusters based on key frames fast recognition algorithm that can quickly and accurately identify the target scene, and support the scene changing and system real-time operation.3) We build a common real-time AR systems and 3D rendering platform on mobile for the integration of real-time and the reality provided a guarantee.