Binocular Ranging Accuracy via Confidence-Based Disparity Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional binocular ranging algorithms face challenges in achieving accurate distance estimation due to low confidence levels and matching errors in disparity maps, particularly in scenarios with obstructions or repetitive textures, leading to inaccurate distance calculations.
Innovation Solution
The method involves extracting features from left and right images using a block matching algorithm, normalizing disparity values to obtain a confidence volume, and generating a confidence map to filter out low-confidence pixels, resulting in a target disparity map that enhances the accuracy of distance estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a complete dense disparity map is generated using conventional block matching algorithms, then the disparity map covers all pixel points, but the distance estimation accuracy deteriorates due to inclusion of low confidence pixel points with matching errors
Solution Approach 1:
The patent segments the disparity map generation process by treating high confidence and low confidence pixel points differently. It generates a complete dense disparity map but then segments the pixel points based on confidence levels, applying different handling (retention vs. filtering) to different segments. This allows maintaining completeness while improving accuracy by excluding low confidence points from distance estimation.
Solution Approach 2:
The patent extracts and identifies low confidence pixel points from the complete dense disparity map using confidence level thresholds. These extracted low confidence points are then separated and excluded from the final distance estimation process, removing the harmful elements that would otherwise degrade measurement precision.
2Measurement precision
If all pixel points in the disparity map are used for distance estimation, then the calculation is simple and fast, but the accuracy deteriorates due to matching errors in low confidence regions
Solution Approach 1:
The patent performs preliminary action by generating confidence levels for all pixel points and identifying low confidence regions before the final distance estimation. This preliminary classification allows the system to pre-mark which points should be excluded, so that when distance estimation is performed, only high confidence points are considered, improving accuracy without adding significant complexity to the main calculation flow.
3Measurement precision
If a filter is applied to remove low confidence pixel points, then the distance estimation accuracy improves, but the processing time and computational complexity increase
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different regions in the disparity map based on their confidence levels. Instead of applying a uniform filtering approach to the entire map, it locally identifies and marks only the low confidence pixel points for exclusion. This localized approach minimizes the processing overhead compared to global filtering methods, as it only requires confidence level calculation and threshold comparison rather than complex filtering operations.
Data Source
AI summary
The present disclosure provides a method and an apparatus for binocular ranging, capable of achieving an improved accuracy of binocular ranging. The method includes: extracting features from a left image and a right image to obtain a left feature image and a right feature image; selecting a standard feature image and obtaining a cost volume of the standard feature image by applying a correlation calculation to the left feature image and the right feature image using a block matching algorithm; obtaining a confidence volume by normalizing computational costs of all disparity values in a disparity dimension for each pixel point in the cost volume; obtaining a confidence map by selecting a maximum value from confidence levels of all the disparity values in the disparity dimension for each pixel point in the confidence volume; obtaining a mask map by mapping each pixel point having a confidence level higher than a predetermined threshold in the confidence map to 1 and mapping each pixel point having a confidence level lower than or equal to the threshold in the confidence map to 0; obtaining a disparity map by calculating an argmax value for the confidence levels of all disparity values in the disparity dimension for each pixel point in the confidence volume; obtaining a target disparity map by multiplying the mask map with the disparity map; and estimating a distance based on the target disparity map.