Traffic Light Detection Area Adjustment via Vehicle Behavior
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional traffic light recognition systems face challenges in accurately setting the detection area due to a time difference between self-position estimation updates and image capture times, especially when vehicle behavior is significant, leading to misalignment and ineffective detection.
Innovation Solution
The system calculates the movement amount of the vehicle during the time difference using vehicle behavior information and adjusts the detection area accordingly, ensuring accurate positioning of the traffic light within the image by estimating the self-position at the time of image capture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the detection area is set based on map information and self-position estimation without considering time difference, then the detection area can be set using available data, but the detection area will be positioned backward to the past and traffic light detection accuracy deteriorates
Solution Approach 1:
The system performs preliminary calculation of vehicle movement amount based on vehicle behavior information (speed, steering angle) before setting the detection area. This allows the detection area to be proactively adjusted to the predicted vehicle position at image capture time, rather than reactively adjusting after the time difference occurs. The detection area is set in advance considering the predicted displacement, ensuring accurate traffic light detection despite timing differences between self-position estimation updates and image capture.
2Measurement precision
If the self-position estimation update frequency is increased to reduce time difference, then detection accuracy improves, but system complexity and processing load increase
Solution Approach 1:
The system introduces vehicle behavior information (speed, steering angle, acceleration) as an intermediary parameter to bridge the time difference gap. Instead of increasing self-position estimation update frequency, the system uses vehicle behavior data to calculate and compensate for the displacement caused by the time difference. This intermediary approach allows accurate detection area positioning without increasing system complexity or processing load.
3Measurement precision
If the detection area is adjusted based on vehicle movement, then traffic light detection accuracy improves, but the complexity of area calculation increases
Solution Approach 1:
The system changes the parameters used for detection area calculation by incorporating vehicle behavior information (speed, steering angle, acceleration) into the area positioning calculation. Instead of using only static map information and self-position estimation, the system dynamically adjusts the detection area parameters based on calculated vehicle displacement. This parameter change approach maintains relatively simple calculation processes while achieving accurate detection area positioning.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A traffic light recognition device includes: a camera (11) that captures an image around a vehicle (51); a map information acquisition unit (17) configured to acquire map information around the vehicle; a self-position detector (12) configured to detect the position of a self-position on a map of the vehicle; a traffic light position estimator (13) configured to estimate the position on an image of a traffic light on the basis of the self-position and the map information; a vehicle behavior estimator (14) configured to estimate a behavior of the vehicle; a traffic light detection area calculator (15) configured to set a detection area of the traffic light onto the image on the basis of the estimated position on the image of the traffic light and an amount of displacement of the position on the image of the traffic light due to a behavior of the vehicle; and a traffic light recognition unit (16) configured to detect a traffic light from the detection area. Such configuring allows the detection area to be set at a suitable position.