Method and device for calibrating model parameters of intelligent terminal and its video image stabilization system

Abstract

The invention discloses an intelligent terminal and a video image stabilization system model parameter calibration method and device thereof. The method comprises the following steps: obtaining a motion information practical reference value; obtaining a motion information estimated value according to the current model parameter; judging whether the motion information estimated value is matched with the motion information practical reference value; and if not, adjusting the model parameters by adopting a multivariable function optimization algorithm with one or more model parameters as variables. Only the newest interframe motion estimation result is utilized, and iterative optimization of the model parameters is carried out by utilizing the multivariable function optimization algorithm; according to the method, all of the previous interframe motion estimation results do not need to be subjected to recomputation and matching rate summarizing according to the adjusted parameters, thereby greatly reducing algorithm complexity; and the model parameter calibration process can be finished by a camera device independently, so that on-line calibration of the model parameters of terminal equipment after leaving factory is realized, a manufacturer no longer needs to carrying out calibration on the terminal equipment before leaving the factory, and production cost is reduced.

Description

technical field [0001] The invention relates to the field of electronic information technology, in particular to a method and device for calibrating model parameters of an intelligent terminal and its video image stabilization system. Background technique [0002] During the recording process of the video recording device, the shaking of the device itself will cause corresponding shaking in the original video it collects. Especially for some portable handheld mobile terminals with a video recording function, the shaking range is relatively large when taking pictures, so the original video collected is often accompanied by obvious picture shaking. [0003] In recent years, with the maturity of the motion sensor, the video recording device can use the motion sensor to sense the shaking of the device, so as to stabilize and stabilize the video image. [0004] Video stabilization solutions based on motion sensors can be divided into optical image stabilization solutions and dig...

AI Technical Summary

Problems solved by technology

But in fact, for each device, the optimal values ​​of these model parameters are not the same, and the use of non-optimal model parameters will affect the effect of video stabilization;

[0010] 3) These parameter calibration requirements increase the cost of mass production and put forward additional requirements for the overall system design of the camera

[0015] The above-mentioned process of obtaining the global optimal value involves relatively large computational complexity, which needs to be completed offline with the help of a computer, and cannot be performed independently by the video recording device itself.

Therefore, this calibration method based on global motion estimation cannot be directly applied to the online calibration of model parameters after leaving the factory.

Images