Video tag dynamic superposition anti-shake method and system based on AR (Augmented Reality)

By constructing a dynamic perspective projection model and a local feature search window, the pixel displacement of AR tags is calculated and compensated, solving the drift and jitter problems of AR tags in complex environments, and achieving a high-precision video overlay effect with low computational overhead.

CN122293995APending Publication Date: 2026-06-26广西信安锐达科技有限公司
0 Cites 0 Cited by

Patent Information

Application Number
CN202610356827.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-23
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

In existing technologies, AR tags suffer from problems such as large cumulative drift error, high computing power consumption, and high-frequency jitter caused by minor vibrations of the device in complex environments, resulting in insufficient visual stability of video overlay.

Method used

By acquiring the real-time video stream and device pose parameters of the front-end monitoring device, a dynamic perspective projection model is constructed, the initial pixel coordinates of the target AR tag in the current video frame are calculated, a local feature search window is planned, the average pixel displacement vector of the image feature points is calculated, and reverse displacement compensation is performed to render the AR tag.

Benefits of technology

It improves the accuracy of label mapping, reduces system computing power overhead, enhances the visual stability of video overlay, and overcomes drift errors caused by mechanical pose dependence and jitter caused by equipment vibration.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a video tag dynamic superposition anti-shake method and system based on AR. The method comprises the following steps: acquiring a real-time video stream and an equipment pose parameter under a corresponding timestamp; constructing a dynamic perspective projection model according to the equipment pose parameter, and calculating an initial pixel coordinate of the target AR tag in the current video frame; planning a local feature search window according to the initial pixel coordinates; calculating an average pixel displacement vector of the image feature points of the current and adjacent historical video frames in the local feature search window; and performing reverse displacement compensation on the average pixel displacement vector to obtain a target rendering coordinate, and rendering the target AR tag to the current video frame based on the target rendering coordinate. Compared with the prior art, the method has the advantages that the defects of large drift error caused by pure dependence on mechanical poses, high computing power consumption caused by global matching and high-frequency jitter of labels caused by small vibration of equipment can be overcome, the label mapping accuracy is improved, the computing power overhead of a system is reduced, and the visual stability of video superposition is improved.
Need to check novelty before this filing date? Find Prior Art