Video Stabilization via Moving Object Feature Point Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic video stabilization methods based on algorithms are inefficient for real-time applications due to high operation costs, limited flexibility, large operation amounts, and slow processing speeds.
Innovation Solution
A system and method for video stabilization that involves obtaining a target frame, detecting feature points, removing points related to moving objects, and performing stabilization based on remaining feature points, while adding supplement feature points to maintain a consistent number of tracked points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing electronic video stabilization methods based on algorithms are used, then video stabilization can be achieved, but the operation amount is large and processing speed is slow
Solution Approach 1:
The patent extracts and removes feature points corresponding to moving objects from the set of feature points used for stabilization calculation. By separating moving object feature points from static background feature points, the system reduces the operation amount in motion parameter estimation while maintaining accurate stabilization of the static scene, thus resolving the contradiction between stabilization reliability and processing speed.
2Reliability
If existing electronic video stabilization methods based on algorithms are used, then video stabilization can be achieved, but the operation amount is large
Solution Approach 1:
The patent identifies and extracts feature points that correspond to moving objects, then removes them from the feature point set used for calculating motion parameters. This extraction process reduces the number of feature points that need to be processed in subsequent stabilization calculations, thereby reducing the overall operation amount and computational complexity while maintaining accurate stabilization performance.
Solution Approach 2:
The patent segments the feature points into two categories: those corresponding to moving objects and those corresponding to static background. By making this segmentation, the system can selectively use only the static background feature points for stabilization parameter estimation, reducing the operation amount without compromising the reliability of video stabilization.
3Quantity of substance
If feature points from moving objects are included in stabilization calculation, then more feature points are available, but the stabilization accuracy decreases
Solution Approach 1:
The patent removes feature points corresponding to moving objects from the feature point set before performing stabilization calculations. Although this reduces the total number of feature points available, it significantly improves stabilization accuracy by eliminating the distorting influence of moving objects on the motion parameter estimation, thereby resolving the contradiction between quantity and precision.
Data Source
AI summary
A method for video stabilization may include obtaining a target frame of a video; obtaining a plurality of first feature points associated with the target frame; determining whether the plurality of first feature points include at least one feature point relating to a moving object in the video; in response to a determination that the plurality of first feature points include at least one feature point relating to the moving object in the video, removing the at least one feature point relating to the moving object; performing video stabilization to the target frame based on remaining first feature points of the plurality of first feature points; determining, in the target frame, at least one supplement feature point; and designating the at least one supplement feature point and the remaining first feature points as second feature points associated with a frame immediately following the target frame in the video.