Research On Camera Locating And Tracking Algorithm In Augmented Reality

AR(Augmented Reality)technology is popular,which is a kind of image and video synthesis technology,it can combine real environment with virtual through projecting images to enhance one's current perception and real environment.It is widely applied in many fields such as the military training,the education,the movie business and other entertainments.The precise locating and real-time tracking of the camera is one of the core technology steps,which is to obtain the six degrees of freedom of the camera,including three parameters of the rotation and translations.Camera locating and tracking methods are mainly used for camera pose estimation based on single point feature.This method can achieve good effects under the feature points rich scene,but the feature points are vulnerable influence to light,shade,and the fast moving camera,resulting in feature points matching error problems.In addition,due to the 2D QR codes are widely used and it has a wealth of features such as points and lines.Thus using the QR code as the identification mark,an algorithm is proposed to locate and track the camera based on the hybrid features.Using the QR code as the identification mark of the augmented reality system,this paper proposes a two-dimensional QR codes borders virtual line extraction method.And this method is based on the hybrid features to a extent to avoid failing of only feature points for positioning and tracking.The experimental results show that the method of combining straight lines with feature points is better than that of using only straight lines or feature points.Further,we also develop an augmented reality demo.