Stereo Matching Disparity Map Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Stereo matching technologies face challenges in accurately calculating distance based on disparity between images from different viewpoints, leading to inaccuracies in distance calculation and user interactions, particularly due to camera motion and dynamic objects.
Innovation Solution
A processor-implemented method that generates a disparity map by transforming a previous frame based on camera motion information, calculates confidence in the disparity map, and adjusts the disparity map using confidence and cost volume aggregation to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a disparity map is generated using traditional stereo matching methods, then the distance calculation can be performed, but the accuracy is reduced due to camera motion and dynamic objects
Solution Approach 1:
The patent applies dynamics by transforming the previous frame's disparity map using estimated camera motion parameters (rotation and translation) to adapt it to the current frame's viewpoint. This dynamic transformation allows the system to maintain accurate depth information even when the camera moves between frames, resolving the contradiction between measurement precision and reliability under motion conditions
Solution Approach 2:
The patent performs preliminary action by estimating camera motion parameters and transforming the previous disparity map before performing the actual stereo matching on the current frame. This preliminary transformation prepares the reference data in advance, allowing the system to achieve high accuracy in distance calculation while accounting for camera motion effects
2Productivity
If the previous frame's disparity map is directly used for the current frame, then processing speed is improved, but accuracy deteriorates due to camera motion
Solution Approach 1:
The system dynamically transforms the previous disparity map using estimated camera motion parameters (rotation matrix and translation vector) to compensate for viewpoint changes. This dynamic adaptation maintains measurement precision while preserving the efficiency benefits of reusing previous frame data, thus resolving the contradiction between productivity and measurement precision
Solution Approach 2:
The patent changes the parameters of the previous disparity map by applying geometric transformation based on camera motion parameters. The disparity values are adjusted according to the rotation and translation between frames, allowing the system to maintain high accuracy while efficiently reusing computational results from the previous frame
3Measurement precision
If confidence calculation is performed to filter inaccurate information, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent implements feedback by calculating confidence values that indicate the reliability of each disparity value in the transformed disparity map. This confidence information is then used to filter or adjust the disparity values before final distance calculation, creating a feedback loop that improves measurement precision while managing processing complexity through selective refinement
Data Source
AI summary
Provided is a stereo matching method and apparatus. A stereo matching apparatus may generate a disparity map by transforming a disparity map of a previous frame based on determined motion information of a camera between the previous frame and the current frame, calculate a confidence for the generated disparity map, and adjust a disparity map corresponding to the current frame based on the confidence and the generated disparity map.