Easy augmented reality

This paper introduces the design and implementation process of an AR project, aiming to embed virtual objects into videos. Firstly, the world coordinate system is fixed and the three-dimensional world coordinates of the object to be rendered are set. Then, camera intrinsic calibration is completed, including distortion parameters. Afterwards, the PnP algorithm is used to solve the rotation and translation relationship between the world coordinate system and the camera coordinate system in each frame, projecting the three-dimensional coordinates of the rendering object onto the image. This paper focuses on the implementation of the PnP algorithm, calculating the relationship between the world coordinate system and the camera coordinate system through feature point tracking and PnP calculation. The algorithm process is divided into four steps, including initial pose estimation, cube projection, optical flow tracking, and continuous PnP pose calculation. In actual operation, there are sources of error, including intrinsic errors, world coordinate point errors, and instability in PnP calculations. Overall, by combining the PnP algorithm with real-time optical flow tracking technology, the project achieves the goal of dynamically inserting virtual objects into videos, providing a new method for the application of augmented reality technology.