Stereo-Camera Object Detection with Adaptive Disparity-Map Height Checks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object detection devices in mobile objects, such as automobiles, face challenges in improving detection performance, particularly in accurately determining the height of objects using stereo cameras.
Innovation Solution
The object detection device employs a processor to search for a target disparity in a disparity map, updating coordinates and calculating object heights based on these disparities, distinguishing between candidate heights using disparities associated with different coordinates in the map to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object detection is performed using disparity maps from stereo cameras, then detection capability is improved, but false positives from road surface structures and parallel objects increase
Solution Approach 1:
The patent segments the detection process into multiple stages: initial object candidate extraction from disparity maps, classification into object/road surface/parallel object categories, and selective height calculation only for validated object candidates. This segmentation isolates false positive sources and applies different processing strategies to each category, reducing overall false positives while maintaining detection accuracy.
Solution Approach 2:
The patent introduces an intermediary classification step that acts as a mediator between raw disparity map data and final object detection results. This classification layer identifies and filters out road surface structures and parallel objects before they can be misclassified as target objects, thereby reducing false positives without compromising genuine object detection.
2Measurement precision
If multiple candidate coordinates are evaluated for object height calculation, then detection accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent implements dynamic coordinate selection where the number of candidate coordinates evaluated (1 or 2) is determined adaptively based on the classification result. For parallel objects, only one coordinate is used; for genuine objects, two coordinates are evaluated. This dynamic approach optimizes processing complexity while maintaining measurement precision through selective evaluation.
Solution Approach 2:
The patent changes the parameter of coordinate evaluation count based on object type classification. By switching between evaluating one coordinate (for parallel objects) and two coordinates (for genuine objects), the system adjusts processing intensity dynamically, reducing overall computational complexity while preserving accuracy for critical detections.
Data Source
AI summary
A processor of an object detection device is configured to search for a pair of coordinates associated with a target disparity in a second direction of a disparity map, update a first pair of coordinates to the found pair of coordinates, and calculate a height of an object corresponding to the target disparity on the basis of the first pair of coordinates. The processor is configured to, when a pair of candidate coordinates associated with a disparity approximately equal to the target disparity is present beyond a predetermined interval from the first pair of coordinates, calculate a first candidate height on the basis of the first pair of coordinates and calculate a second candidate height on the basis of the pair of candidate coordinates.