Traffic Light Detection Using Low-Res Screening and High-Res Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating due to the vast amount of data they need to process and store, particularly with traditional mapping technologies, which can limit their ability to efficiently analyze and update maps while traveling.
Innovation Solution
The use of multiple cameras to provide navigation features, where a processor analyzes images to identify traffic lights and determine their state, allowing the vehicle to take appropriate navigational actions based on confirmed traffic light states, and the implementation of a sparse map for autonomous vehicle navigation that requires less data storage and transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional mapping technology is used for autonomous vehicle navigation, then comprehensive map data can be provided, but the data volume becomes excessively large making storage and updates difficult
Solution Approach 1:
The patent extracts only the essential navigation elements from comprehensive map data, creating a sparse map that contains only critical information needed for autonomous navigation. This reduces data volume while maintaining navigation reliability by focusing on key features rather than storing complete traditional maps.
Solution Approach 2:
The patent implements local quality by providing high-resolution detailed data only where needed for navigation decisions, while using lower-resolution or aggregated data for other areas. This allows the system to maintain navigation accuracy in critical zones without storing excessive data throughout the entire map.
2Measurement precision
If high resolution image analysis is performed for traffic light detection, then detection accuracy is improved, but processing time and computational burden increase
Solution Approach 1:
The patent segments the image analysis process into two stages: first analyzing a downsampled low-resolution image to identify candidate traffic light regions, then performing high-resolution analysis only on those specific candidate regions. This segmentation reduces overall processing time while maintaining detection accuracy for actual traffic lights.
Solution Approach 2:
The patent applies partial action by performing high-resolution analysis only on candidate regions rather than the entire image. This selective approach processes only the necessary portions at high resolution, reducing computational burden while ensuring accurate detection of traffic lights when they are present.
Data Source
AI summary
Systems and methods are provided for vehicle navigation. The systems and methods may detect traffic lights. For example, one or more traffic lights may be detected using detection-redundant camera detection paths, a fusion of information from a traffic light transmitter and one or more cameras, based on contrast enhancement for night images, and based on low resolution traffic light candidate identification followed by high resolution candidate analysis. Additionally, the systems and methods may navigation based on a worst time to red estimation.