Binocular Image Phase Difference Estimation via Multi-Level Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing binocular image processing methods face challenges in accurately acquiring phase differences due to disparities in image sizes and lower accuracy, leading to inefficient image processing.
Innovation Solution
A binocular image quick processing method and apparatus that employs folding dimensionality reduction, feature extraction using residual convolutional networks, phase difference distribution estimation, fusion of features, and tiling dimensionality raising to accurately estimate phase differences across multiple levels of image resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional binocular image processing methods are used, then the processing speed may be maintained, but the accuracy of phase difference acquisition deteriorates due to image size disparities and feature precision differences
Solution Approach 1:
The patent segments the binocular image processing into multiple levels (first-level, second-level, third-level images) with different resolutions. Each level processes features at its own scale, allowing accurate phase difference measurement across different object sizes without overwhelming computational complexity at any single stage.
Solution Approach 2:
The patent introduces a multi-resolution dimension by creating images at different levels (first-level, second-level, third-level). This dimensional approach allows the system to analyze features at appropriate scales for different object sizes, improving phase difference accuracy without requiring a single complex processing path.
2Measurement precision
If multi-level processing is implemented to improve accuracy, then phase difference measurement precision improves, but processing time increases
Solution Approach 1:
The patent performs preliminary processing at lower resolutions (second-level, third-level images) before final high-resolution processing. By extracting features and estimating phase differences at reduced resolutions first, the system reduces the computational burden on high-resolution images, maintaining accuracy while reducing overall processing time.
Solution Approach 2:
The patent applies partial processing to different image levels - using lower-resolution images for initial feature extraction and phase difference estimation, then applying more detailed processing only where needed. This selective approach maintains accuracy for critical measurements while reducing unnecessary computational overhead.
3Measurement precision
If uniform processing is applied to all image regions, then processing simplicity is maintained, but accuracy deteriorates for objects of different sizes due to feature precision differences
Solution Approach 1:
The patent applies different processing qualities to different image levels - using lower-resolution processing for regions with smaller objects and higher-resolution processing for regions with larger objects. This local adaptation ensures that feature extraction accuracy matches the scale of objects in each region, improving overall measurement precision.
Solution Approach 2:
The patent changes processing parameters (resolution level, feature extraction depth) based on the scale of objects being analyzed. By adjusting these parameters according to object size and image level, the system maintains consistent feature precision across objects of different sizes without requiring a single uniform complex processing approach.
Data Source
AI summary
The present invention provides a binocular image quick processing method, including: performing feature extraction on a next-level left eye image and a next-level right eye image; acquiring a next-level image phase difference distribution estimation feature; fusing the next-level image phase difference distribution estimation feature and a next-level left eye image feature to obtain a next-level fusion feature; performing feature extraction on the next-level fusion feature to obtain a difference feature of next-level left and right eye images, and obtain an estimated phase difference of the next-level left and right eye images; acquiring an estimated phase difference of first-level left and right eye images; and performing processing operation on the corresponding images by using the estimated phase difference of the first-level left and right eye images.