Traffic Light Recognition Using Map-Image Cross Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic light recognition methods for driverless vehicles often result in inaccurate detection or missed traffic lights due to errors in determining the region of interest, leading to incorrect lane and color identification.
Innovation Solution
A traffic light recognition method that uses a combination of image detection and map positioning data, with a preset error constraint to match traffic lights, ensuring accurate location and indication information matching between image and map data, thereby reducing recognition errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If coordinate transformation is performed on location information of traffic light group to determine region of interest, then the traffic light recognition process can be completed, but errors may exist in determining the region of interest leading to mistaken detection or missed traffic lights
Solution Approach 1:
The patent combines map positioning results with image detection results to determine traffic light location and status. By merging these two independent detection methods, the system achieves both efficient processing and high accuracy, resolving the contradiction between productivity and measurement precision.
Solution Approach 2:
The system uses map positioning results to guide image detection and vice versa, creating a feedback mechanism that continuously refines traffic light detection accuracy while maintaining processing efficiency.
2Device complexity
If only image detection method is used to detect traffic lights, then the detection process is simple, but traffic lights may be mistakenly detected or missed
Solution Approach 1:
The patent merges map positioning results with image detection results to achieve reliable traffic light detection. This combination maintains relatively simple processing while significantly improving detection reliability by cross-validating results from two different methods.
3Ease of operation
If only map positioning information is used to determine traffic light location, then the process is straightforward, but errors beyond expectation lead to incorrect lane and color determination
Solution Approach 1:
The system implements feedback by using image detection results to verify and correct map positioning results. This ensures both operational simplicity and high location precision by leveraging the strengths of both methods.
Solution Approach 2:
The patent uses an error constraint mechanism as an intermediary to reconcile differences between map positioning and image detection results, ensuring accurate traffic light identification while maintaining process simplicity.
Data Source
AI summary
This application discloses a traffic light recognition method and apparatus in the field of artificial intelligence that relate to an automated driving technology. The recognition apparatus determines location information of n first traffic lights in a first image based on current positioning information of a vehicle, a calibration parameter of an image sensor, and prestored map data; and detects the first image by using an image detection method, to obtain location information of m second traffic lights in the first image and indication information of the m second traffic lights. This effectively reduces a traffic light recognition error, and improves recognition accuracy.