Video Stabilization via Static Region Motion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional algorithm-based video stabilization methods are hindered by the presence of moving objects in images, leading to interference in motion estimation and reduced video quality.
Innovation Solution
The method involves selecting a first area in an image with minimal motion, determining feature points, conducting feature point matching between images, estimating motion parameters, and performing motion compensation to stabilize the image, thereby avoiding interference from moving objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If algorithm-based video stabilization method is used to perform motion estimation on all objects in an image, then motion deviation can be grasped and compensated, but moving objects in the image interfere with the calculation result and reduce video quality
Solution Approach 1:
The patent divides the image into multiple regions of interest (ROIs) and selectively performs motion estimation on specific regions rather than the entire image. This segmentation allows the system to focus on static background areas while excluding moving objects from the motion estimation process, thereby reducing interference and improving stabilization reliability.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image. By identifying and isolating static background regions from moving objects, the system applies motion estimation and compensation only to appropriate regions, ensuring that moving objects do not adversely affect the overall stabilization calculation.
2Measurement precision
If motion estimation is performed on the entire image including moving objects, then comprehensive motion data is obtained, but excess compensation occurs and video quality deteriorates
Solution Approach 1:
The patent extracts and removes moving objects from the set of objects used for motion estimation. By identifying moving objects and excluding them from the motion estimation process, the system obtains more precise motion data from static background regions, thereby improving both measurement precision and the resulting motion compensation accuracy.
Solution Approach 2:
Instead of performing motion estimation on all objects in the image (which would include moving objects causing excess compensation), the patent applies motion estimation partially to only the static background regions. This partial action approach ensures that motion compensation is based on accurate data from appropriate sources, avoiding the excessive compensation that would result from including moving objects.
Data Source
AI summary
Disclosed is a video stabilization method including steps of selecting a first area in a first image; determining at least one first feature point based on the first area, and performing feature point matching on the first image and a second image so as to obtain at least one feature point matching pair between the first image and the second image, each feature point matching pair including a first feature point of the first image and a second feature point of the second image; conducting motion parameter estimation based on the at least one feature point matching pair so as to obtain a motion parameter; and carrying out motion compensation with respect to the first image based on the motion parameter so as to acquire a first stable image after compensation.