Stereoscopic Ground-Plane Discrimination Using Expected Disparity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Stereoscopic vision systems face computational inefficiencies in feature extraction and matching, limiting their application in tasks requiring 3D information processing.
Innovation Solution
A method for rapidly discriminating between on-ground-plane and off-ground-plane locations using expected disparity comparisons in stereoscopic images, allowing for computationally inexpensive processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature extraction and feature matching are performed in stereoscopic vision systems, then 3D information can be extracted, but computational cost increases significantly
Solution Approach 1:
The patent extracts only the essential disparity information needed for ground-plane discrimination, rather than performing full feature extraction and matching. By focusing solely on comparing image portions at expected disparity locations, the system achieves 3D spatial understanding without the computational burden of comprehensive feature analysis.
Solution Approach 2:
The patent segments the stereoscopic images into discrete image portions that can be independently compared. By dividing the images and processing them in manageable segments with simple pixel-level comparisons, the system avoids the need for complex global feature extraction while maintaining effective ground-plane discrimination.
2Measurement precision
If feature extraction and matching are performed, then 3D information is obtained, but processing time increases
Solution Approach 1:
The patent uses pre-calculated expected disparity values to directly locate corresponding image portions without requiring time-consuming feature detection and matching. By performing the disparity calculation in advance and using it to guide the comparison process, the system significantly reduces processing time while maintaining accurate 3D spatial discrimination.
Solution Approach 2:
The patent skips the intermediate steps of feature detection, feature description, and feature matching that characterize traditional stereoscopic vision systems. By rushing directly to comparing image portions at expected disparity locations, the system achieves rapid ground-plane discrimination with processing time reduced from hundreds of milliseconds to minimal delays.
3Productivity
If simple comparison methods are used, then processing speed increases, but discrimination accuracy may decrease
Solution Approach 1:
The patent introduces expected disparity values as an intermediary that guides the simple pixel-level comparison process. This intermediary provides the theoretical foundation for determining corresponding image portions, ensuring that even simple comparisons yield accurate ground-plane discrimination results by comparing the right image portions with each other.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A stereoscopic vision system captures stereoscopic images. The stereoscopic images are processed to rapidly discriminate between portions of the images that are on a ground-plane and those that are off the ground plane. The discrimination is based on comparing locations within the images using an expected disparity shift.