Camera tracking method and device

Abstract

Provided are a camera tracking method and device, which use a binocular video image to perform camera tracking, thereby improving the tracking accuracy. The camera tracking method provided in the embodiments of the present invention comprises: acquiring an image set of a current frame; respectively extracting feature points of each image in the image set of the current frame; according to a principle that depths of scene in adjacent regions on an image are similar, acquiring a matched feature point set of the image set of the current frame; according to an attribute parameter and a pre-set model of a binocular camera, respectively estimating three-dimensional positions of scene points corresponding to each pair of matched feature points in a local coordinate system of the current frame and a local coordinate system of the next frame; and according to the three-dimensional positions of the scene points corresponding to the matched feature points in the local coordinate system of the current frame and the local coordinate system of the next frame, estimating a motion parameter of the binocular camera in the next frame using the invariance of a barycentric coordinate with respect to rigid transformation, and optimizing the motion parameter of the binocular camera in the next frame.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a camera tracking method and device. Background technique [0002] Camera tracking (Camera tracking) is one of the most basic problems in the field of computer vision. According to the video sequence shot by the camera, the three-dimensional position of the feature point in the shooting scene and the camera motion parameters corresponding to each frame of image are estimated; with the rapid development of technology With the progress, the application fields of camera tracking technology are very extensive, such as robot navigation, intelligent positioning, combination of virtual reality, augmented reality, three-dimensional scene browsing, etc.; in order to adapt to the application of camera tracking in various fields, after decades of hard research, some cameras Tracking systems have also been launched, such as PTAM (Parallel Tracking and Mapping), ACTS (Automatic Camera Tracking ...

AI Technical Summary

Problems solved by technology

[0003] In practical applications, PTAM and ACTS systems perform camera tracking for monocular video sequences. During the camera tracking process, two frames need to be selected as initial frames. figure 1 It is a schematic diagram of camera tracking based on monocular video sequences in the prior art, such as figure 1 As shown, using the matching points of the initial frame 1 image and the initial frame 2 image (x 1,1 ,x 1,2 ) to estimate the relative position between the corresponding cameras of the two initial frame images (R 12 ,t 12 ); Initialize matching points by triangulation (x 1,1 ,x 1,2 ) corresponds to scene point X 1 3D position; when tracking the subsequent frame, use the corresponding relationship between the known 3D point position and the 2D point in the subsequent frame image to solve the camera motion parameters of the subsequent frame; however, the camera initialized in the camera tracking based on the monocular video sequence The relative position between (R 12 ,t 12 ) estimates have errors, and these errors are transmitted to the estimation of subsequent frames through the uncertainty of the scene, making the errors accumulate in the tracking of subsequent frames, which are difficult to eliminate, and the tracking accuracy is low

Images