Road Area Detection Learning Device Multi-Viewpoint Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for detecting road areas in images for vehicle driving support or automatic driving lack verification from multiple viewpoints, leading to potential inaccuracies in road area detection.
Innovation Solution
A learning device and method that trains a machine learning model to output pixel regions representing road edges and areas, with a determining unit comparing these regions to determine matching degrees, a setting unit adjusting error gains based on matching degrees, and a learning unit reducing a loss function based on these errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If road area detection is performed using single-viewpoint scanning technology, then the detection process is simple, but the detection accuracy cannot be ensured
Solution Approach 1:
The patent introduces a multi-viewpoint detection dimension by capturing images from multiple cameras positioned at different locations (front, rear, left, right). This spatial dimensionality change allows verification of road area detection results from multiple perspectives, significantly improving detection accuracy while maintaining reasonable system complexity through modular camera arrangement.
Solution Approach 2:
The patent implements a feedback mechanism where detection results from multiple viewpoints are compared and verified against each other. The determination unit uses feedback from consistent detection results across multiple views to confirm accuracy, and triggers re-detection or correction when inconsistencies are found, thereby ensuring high detection accuracy through iterative verification.
2Reliability
If multiple viewpoints are used to verify road area detection, then detection accuracy is improved, but the complexity of the detection system increases
Solution Approach 1:
The patent designs the multi-viewpoint detection system with universal functionality where each camera serves multiple purposes: primary detection of road areas, verification of detection results, and provision of backup detection data. This multi-functionality approach allows the system to achieve high reliability through multiple viewpoints while managing complexity by having each component perform several functions simultaneously.
Solution Approach 2:
The patent segments the detection system into independent modular units (separate cameras for front, rear, left, right views) that can be individually configured and processed. This segmentation allows the complex multi-viewpoint system to be managed as separate functional modules, reducing overall system complexity while maintaining high reliability through distributed detection and verification capabilities.
Data Source
AI summary
A learning device for training a machine learning model that receives an image as an input and outputs a first pixel region representing road edges and a second pixel region representing a road area in the image includes a storage medium storing computer-readable instructions and a processor connected to the storage medium, the processor executing the computer-readable instructions to compare the first and second pixel regions to determine a degree of matching between the first and second pixel regions, set a gain for a first error between the first pixel region and correct data representing the road edges and a second error between the second pixel region and correct data representing the road area based on the degree of matching, and train the machine learning model so as to reduce a value of a loss function calculated based on the first and second errors for which the gain has been set.