Stereo Matching Search Offset Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In three-dimensional display technologies, analyzing two-dimensional images to obtain depth information requires a full-range matching search between images with different views, leading to increased computing costs, time, and power consumption, especially with high-definition images.
Innovation Solution
A matching search method and system that dynamically determines a search position and range by calculating a global offset between image blocks, allowing the search to be performed within a predefined smaller range without compromising accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full-range matching search is performed within the entire data range of two images to obtain accurate matching results, then the accuracy of matching search results is improved, but the computing cost, computing time, and power consumption are significantly increased
Solution Approach 1:
The patent performs a preliminary full-range matching search to obtain initial matching results and calculate a global offset. This preliminary action establishes a reference frame that enables subsequent restricted-range searches to achieve accurate results without examining the entire image range, thus reducing computing cost while maintaining accuracy.
Solution Approach 2:
The patent divides the matching search into different ranges with different search depths. A first matching search is performed in a first range with a first search depth, and a second matching search is performed in a second range with a second search depth. This local differentiation allows the system to allocate computing resources efficiently, focusing computational effort where it is most needed while reducing effort in regions that can be searched more lightly.
2Measurement precision
If a full-range matching search is performed within the entire data range of two images to obtain accurate matching results, then the accuracy of matching search results is improved, but the computing time is significantly increased
Solution Approach 1:
The patent performs a preliminary full-range matching search to obtain initial matching results and calculate a global offset. This preliminary action establishes a reference frame that enables subsequent restricted-range searches to achieve accurate results without examining the entire image range, thus reducing computing time while maintaining accuracy.
Solution Approach 2:
The patent dynamically adjusts the search range and search depth based on the global offset calculated from preliminary results. The second matching search uses a second range and second search depth that are adapted from the first search results, allowing the system to optimize computing time by avoiding redundant searches in regions where the global offset indicates low probability of new matches.
3Measurement precision
If a full-range matching search is performed within the entire data range of two images to obtain accurate matching results, then the accuracy of matching search results is improved, but the hardware cost is significantly increased
Solution Approach 1:
The patent performs a preliminary full-range matching search to obtain initial matching results and calculate a global offset. This preliminary action establishes a reference frame that enables subsequent restricted-range searches to achieve accurate results without examining the entire image range, thus reducing hardware cost while maintaining accuracy.
Solution Approach 2:
The patent divides the matching search into different ranges with different search depths. A first matching search is performed in a first range with a first search depth, and a second matching search is performed in a second range with a second search depth. This local differentiation allows the system to allocate computational resources efficiently, reducing the hardware power and complexity required compared to a uniform full-range search.
Data Source
AI summary
A matching search method is utilized for performing a matching search from a first image block to a second image block in a stereo matching system. The matching search method includes obtaining a global offset of the first image block corresponding to the second image block according to an offset relationship between a first image data of the first image block and a second image data of the second image block; and performing the matching search from the first image block to the second image block according to the global offset and a searching range.