Stereo Matching Feature Point Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current stereo matching methods face challenges in accurately determining depth information from stereo images, particularly in regions with depth discontinuities, occlusions, and low texture areas, leading to inaccuracies and increased computational resources.
Innovation Solution
A method that extracts feature points from stereo images, classifies reference points based on depth discontinuities, occlusions, and texture reliability, and updates disparity values using a reliability-based propagation approach to enhance accuracy and reduce computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional stereo matching methods are used to determine depth information, then computational resources are consumed, but accuracy deteriorates in regions with depth discontinuities, occlusions, and low texture areas
Solution Approach 1:
The patent segments the stereo matching process into distinct phases: feature point extraction, reference point determination through window-based correlation, classification of reference points into different types (occlusion, depth discontinuity, low texture regions), and selective disparity calculation. This segmentation allows the system to apply different processing strategies to different image regions, improving accuracy while managing computational resources efficiently.
Solution Approach 2:
The patent applies local quality by treating different regions of the stereo image differently based on their characteristics. Reference points are classified into multiple categories (occlusion regions, depth discontinuity regions, low texture regions, and normal regions), and disparity calculation is performed selectively based on these classifications. This ensures high accuracy in challenging regions while avoiding unnecessary computations in straightforward areas.
2Reliability
If feature point matching is performed using window-based correlation with cost analysis, then matching accuracy is improved, but computational time increases
Solution Approach 1:
The patent performs preliminary feature point extraction and reference point determination using window-based correlation and cost analysis before the main disparity calculation process. By pre-identifying and classifying reference points in advance, the system prepares the data structure needed for efficient subsequent processing, reducing the computational burden during the actual stereo matching phase.
Solution Approach 2:
The patent applies partial action by performing exhaustive window-based correlation and cost analysis only on selected reference points rather than all pixels in the image. The classification system identifies which regions require detailed analysis (occlusion, depth discontinuity, low texture areas) and applies the computationally intensive matching only where necessary, rather than uniformly across the entire image.
3Measurement precision
If disparity values are updated based on reliability of reference points, then stereo matching accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent changes the parameter of reference point reliability by introducing a classification system that assigns different reliability levels to different reference points based on their local image characteristics. This parameter change allows the disparity update process to selectively weight or validate disparity values based on the reliability classification, improving accuracy without requiring uniformly complex processing across all regions.
Data Source
AI summary
A stereo matching method includes extracting feature points of a first image and feature points of a second image, the first image and the second image together constituting a stereo image, determining reference points by matching the feature points of the second image to the feature points of the first image, classifying the reference points, and performing stereo matching on pixels of which disparities are not determined in the first image and the second image based on disparities of the reference points in the pixels determined based on a result of the classifying.