Stereo Camera Object Detection Using U-Disparity and V-Disparity Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection systems using stereo cameras often incorrectly detect adjacent objects as a single entity, leading to suboptimal performance in recognizing individual objects.
Innovation Solution
The proposed solution involves an information processing device and method that utilize a combination of parallax images and reference images to generate U-Disparity maps and V-Disparity maps, which aid in accurate object recognition by distinguishing between objects and background.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If object detection is performed using parallax image integration, then detection coverage is improved, but detection precision deteriorates causing adjacent objects to be merged
Solution Approach 1:
The patent divides the detection process into multiple stages: initial detection using integrated parallax images for comprehensive coverage, followed by separation detection that identifies and divides merged objects into individual objects. This segmentation approach resolves the contradiction by first ensuring no objects are missed, then precisely separating adjacent objects that were incorrectly merged.
Solution Approach 2:
The patent introduces a separation detection stage that operates in a different dimensional space by analyzing object boundaries and spatial relationships more carefully. This additional dimensional analysis allows the system to distinguish between truly connected objects and adjacent objects that should be separated, thereby improving precision without sacrificing coverage.
2Productivity
If simple parallax-based detection is used, then processing speed is improved, but detection accuracy deteriorates
Solution Approach 1:
The patent performs preliminary detection using fast parallax-based methods to quickly identify potential objects and their rough locations. This preliminary action provides a good initial result that guides subsequent more accurate but computationally intensive separation detection, thereby maintaining high processing speed while improving accuracy.
Solution Approach 2:
The patent applies partial detailed analysis only where needed - specifically in regions where objects are detected to be adjacent or merged. Rather than performing exhaustive analysis on the entire scene, the system focuses computational resources on problematic areas, achieving high accuracy where it matters most while maintaining overall processing efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the performance of object recognition by improving the accuracy of detecting and separating individual objects from adjacent ones, leading to better decision-making in applications like autonomous vehicles.
Implementation Method 1
a parallax image is generated based on a parallax of each object projected in a taken luminance image
Data Source
Figure 1A~1B
Figure 2
Figure 3
AI summary
An information processing device includes: a first generation unit configured to generate first information in which a horizontal direction position and a depth direction position of an object are associated with each other from information in which a vertical direction position, the horizontal direction position, and the depth direction position of the object are associated with each other; a first detection unit configured to detect one region indicating the object based on the first information; a second generation unit configured to generate, from the information in which the vertical direction position, the horizontal direction position, and the depth direction position of the object are associated with each other, second information having separation performance higher than separation performance of the first information in which the horizontal direction position and the depth direction position of the object are associated with each other; a second detection unit configured to detect a plurality of regions indicating objects based on the second information; and an output unit configured to associate the one region detected based on the first information with the regions detected based on the second information, and to output the one region and the regions that are associated with each other.