Traffic Light Detection Using Candidate Scoring and Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately detecting and classifying traffic lights and other traffic indicators in images captured by onboard cameras, especially under varying environmental conditions, which can impact safe vehicle operation.
Innovation Solution
A method and system that utilize an image processing system equipped with a camera, processor, and memory to calculate confidence scores for image portions, identify candidate portions exceeding a threshold, and classify them using a trained classifier to determine the presence and state of traffic lights, providing instructions for vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional image processing methods are used to detect traffic lights, then the system complexity is low, but the detection accuracy and reliability deteriorate under varying environmental conditions
Solution Approach 1:
The image processing is divided into multiple stages: candidate region identification based on color segmentation, followed by classification using a trained classifier. This multi-stage segmentation approach improves detection reliability by breaking down the complex detection task into manageable steps, each addressing specific aspects of traffic light identification.
Solution Approach 2:
The system performs preliminary action by pre-training a classifier with training data before actual detection. The classifier is trained offline with example images of traffic lights and non-traffic lights, so that during runtime, the system can quickly and accurately classify candidate regions without performing complex real-time training, thus improving reliability while maintaining acceptable system complexity.
2Measurement precision
If a trained classifier is used to classify candidate portions, then the classification accuracy improves, but the processing time and computational resources increase
Solution Approach 1:
The classifier is trained in advance using training data containing example images of traffic lights and non-traffic lights. This preliminary training allows the system to achieve high classification accuracy during runtime without performing computationally intensive training operations in real-time, thus reducing processing time while maintaining precision.
Solution Approach 2:
The system first identifies candidate portions using color-based segmentation, which quickly filters out non-relevant regions. Only these candidate portions are then passed to the trained classifier for detailed classification. This two-stage segmentation approach reduces the number of images requiring complex classification, thereby reducing overall processing time while maintaining high accuracy.
3Reliability
If the system processes all image portions to ensure accurate detection, then the detection completeness improves, but the processing speed deteriorates
Solution Approach 1:
The system segments the image processing task into two phases: first, rapid color-based segmentation identifies candidate regions that may contain traffic lights; second, the trained classifier processes only these candidate portions. This segmentation ensures that no potential traffic light is missed (maintaining detection completeness) while avoiding unnecessary processing of non-relevant image portions (maintaining processing speed).
Solution Approach 2:
The system performs partial action by processing only the candidate portions identified through color segmentation rather than analyzing every pixel in the entire image. This selective processing maintains detection completeness for traffic lights while significantly reducing the overall processing burden, thus improving processing speed without sacrificing reliability.
Data Source
AI summary
The present disclosure is directed to an autonomous vehicle having a vehicle control system. The vehicle control system includes an image processing system. The image processing system receives an image that includes a plurality of image portions. The image processing system also calculates a score for each image portion. The score indicates a level of confidence that a given image portion represents an illuminated component of a traffic light. The image processing system further identifies one or more candidate portions from among the plurality of image portions. Additionally, the image processing system determines that a particular candidate portion represents an illuminated component of a traffic light using a classifier. Further, the image processing system provides instructions to control the autonomous vehicle based on the particular candidate portion representing an illuminated component of a traffic light.