High-Resolution Image Matching via Multi-Level Down-Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current visual depth perception technologies face high computational complexity, limiting them to support only low-resolution image matching and resulting in lower accuracy and depth data for high-resolution images, especially for real-time applications.
Innovation Solution
A high-resolution image matching method involving regional fidelity down-sampling to create multi-level low-resolution images, followed by local matching using global probes and reverse refinement based on overall consistency to improve accuracy and reduce computational volume.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If local pixel-by-pixel matching method is used, then matching accuracy is improved, but computational complexity increases significantly
Solution Approach 1:
The patent divides the high-resolution image into multiple low-resolution images through down-sampling, creating a multi-level resolution pyramid. The matching process is segmented into two stages: first performing matching on low-resolution images to obtain initial correspondence, then refining results on high-resolution images. This segmentation reduces the overall computational burden while maintaining accuracy through progressive refinement.
Solution Approach 2:
The patent performs preliminary matching actions on low-resolution images before processing high-resolution images. By first establishing coarse matching relationships at lower resolutions, the system prepares initial correspondence data that guides subsequent high-resolution matching, avoiding the need to perform exhaustive pixel-by-pixel comparison directly on high-resolution images.
2Productivity
If low-resolution image matching is used to ensure real-time performance, then processing speed is improved, but matching accuracy deteriorates
Solution Approach 1:
The matching process is segmented into multiple resolution levels. Low-resolution matching provides fast initial results, while high-resolution matching refines these results. This segmentation allows the system to benefit from both fast low-resolution processing and accurate high-resolution matching without having to choose one or the other.
Solution Approach 2:
The patent introduces a resolution dimension, creating a multi-level pyramid structure with images at different resolutions. Instead of performing matching in a single resolution space, the system operates across multiple resolution dimensions, transitioning from coarse to fine levels. This dimensional approach enables real-time performance at lower levels while achieving high accuracy at upper levels.
3Measurement precision
If high-resolution image matching is performed directly, then matching accuracy is improved, but computational volume increases excessively
Solution Approach 1:
The computational task is segmented across multiple resolution levels. Rather than computing all matching operations at full resolution, the system segments the work: low-resolution levels handle coarse alignment and reduce search spaces, while high-resolution levels perform detailed refinement only in relevant regions. This segmentation dramatically reduces total computational volume.
Solution Approach 2:
Preliminary matching operations are performed on down-sampled low-resolution images to establish initial correspondence relationships. These preliminary results provide guidance for subsequent high-resolution matching, reducing the search space and computational volume required for the final high-accuracy matching stage.
4Measurement precision
If semi-global optimization method is used, then parallax accuracy is improved, but calculation complexity remains high for high-resolution images
Solution Approach 1:
The optimization process is segmented into multiple resolution levels. Semi-global optimization is applied first at low resolutions where computational complexity is manageable, obtaining optimized parallax results. These results are then used as priors for high-resolution matching, where optimization is performed more efficiently due to the reduced search space provided by lower-level results.
Solution Approach 2:
The patent introduces resolution levels as an additional dimension to the matching process. By organizing images in a resolution pyramid, the system can apply optimization methods at different levels of this dimension, transitioning from coarse optimization at lower resolutions to fine optimization at higher resolutions, thereby managing computational complexity across the dimension of image detail.
Data Source
AI summary
A high-resolution image matching method and system are provided. The method includes performing regional fidelity down-sampling on an initial high-resolution image to obtain a multi-level low-resolution image, performing local matching on the obtained multi-level low-resolution images using a method with global probes to obtain a matching result of the low-resolution images, and performing reverse refinement on the obtained matching result of the low-resolution image using overall consistency of the image matching, to obtain the matching results of the high-resolution images at all levels until the matching results of the initial resolution images are obtained, so as to reduce the computational complexity of the image matching process and improve the accuracy of the matching result, and then the matching result of the high-resolution image is obtained through reverse refinement based on the overall consistency.