Moving Object Detection Using Binned Phase Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional foreground extraction methods fail when dealing with moving backgrounds and foreground objects, as they incorrectly classify everything as foreground due to the assumption of static backgrounds, leading to inaccurate detection of moving objects, especially in scenarios like autonomous vehicles with moving cameras.
Innovation Solution
The method employs binned phase correlation to detect moving objects by comparing images captured at different times, determining affine translations, and transforming images to identify pixels that differ significantly, effectively distinguishing between static and dynamic elements even in the presence of a moving background.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If traditional foreground extraction methods assume static background, then simple motion detection can be performed, but detection accuracy deteriorates when background is moving
Solution Approach 1:
The image is divided into multiple blocks, and each block is processed independently to determine its motion state. This segmentation allows the system to handle different motion characteristics in different regions, improving overall detection accuracy while maintaining computational efficiency
Solution Approach 2:
The system changes the reference parameter from a static background model to a dynamic background model that accounts for camera motion. By using affine transformation parameters to describe background motion and comparing actual block motion against this dynamic reference, the system achieves accurate foreground detection in moving background scenarios
2Measurement precision
If optical flow is used to detect motion differences, then some moving objects can be detected, but motion between sampled points is not detected
Solution Approach 1:
Instead of using optical flow that samples specific points, the patent divides the image into comprehensive blocks that cover the entire image area. This ensures that motion between any points within a block is captured by analyzing the overall block displacement, eliminating the sampling limitation of optical flow
Solution Approach 2:
The patent replaces the complex optical flow computation with a simpler block-matching approach using affine transformation. This substitution reduces computational complexity while maintaining the ability to detect various types of motion, including translation, rotation, and scaling
3Adaptability or versatility
If everything is classified as foreground in moving background scenarios, then no motion filtering is applied, but false positives increase
Solution Approach 1:
The system introduces affine transformation parameters to model background motion and uses these parameters to filter out background movement. By comparing actual block displacement against the modeled background displacement, the system reliably distinguishes foreground objects from background motion, reducing false positives while maintaining adaptability to moving backgrounds
Data Source
AI summary
A method includes identifying a first image that is captured at a first time. The method also includes segmenting the first image into a plurality of first image portions. The method also includes identifying a second image that is captured at a second time. The method also includes segmenting the second image into a plurality of second image portions. The method also includes comparing one of the plurality of first image portions and a corresponding one of the plurality of second image portions. The method also includes determining a difference between the first image and the second image based at least partially upon the comparison. The method also includes transforming the first image into a transformed first image based at least partially upon the difference.