Stereo Camera Object Detection Using Road Surface Disparity Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Object detection devices in mobile objects face challenges in achieving high performance due to road shapes and structures, leading to inaccurate obstacle detection.
Innovation Solution
An object detection system utilizing a stereo camera and an object detection device that generates a disparity map, estimates road surface position, removes unnecessary disparities, and detects objects by converting disparities into x-z coordinates, enhancing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a stereo camera is used to generate a disparity map for object detection, then distance measurement capability is improved, but detection accuracy deteriorates due to road shape and structure interference
Solution Approach 1:
The patent segments the disparity map into road surface regions and non-road surface regions using histogram analysis. By separating the road surface disparity distribution from other regions, the system can independently analyze object disparities without interference from road shape variations, thereby improving detection accuracy while maintaining distance measurement capability
Solution Approach 2:
The patent applies different processing strategies to different regions of the disparity map. Road surface regions are processed using histogram-based disparity extraction, while non-road regions are processed using object detection algorithms. This localized approach allows each region to be handled with the most appropriate method, improving overall detection accuracy
2Area of stationary object
If the entire disparity map is processed for object detection, then detection coverage is improved, but processing time increases
Solution Approach 1:
The patent divides the disparity map processing into two independent streams: road surface disparity extraction through histogram analysis and object detection in non-road regions. This segmentation allows parallel processing of different regions, reducing overall processing time while maintaining comprehensive detection coverage
Solution Approach 2:
The patent applies object detection algorithms selectively to non-road regions identified through histogram analysis, rather than processing the entire disparity map uniformly. This partial action approach reduces computational load and processing time while maintaining detection coverage in critical areas
Data Source
AI summary
An object detection device is configured to execute a road surface detection process, an object disparity determination process, and an object detection process. The road surface detection process estimates a position of a road surface on the basis of a first disparity map. The first disparity map is generated on the basis of an output of a stereo camera and is a map in which a disparity is associated with two-dimensional coordinates formed by a first direction corresponding to a horizontal direction of an image captured by the stereo camera and a second direction intersecting the first direction. The object disparity determination process determines disparities as object disparities when the number of occurrences of each of the disparities for respective coordinate ranges in the first direction of the first disparity map exceeds a predetermined threshold corresponding to the disparity. The object detection process detects an object by converting information on the object disparities into points on an x-z coordinate space and by extracting a group of points.