Hierarchical Traffic Light Detection for Brake Light Distinction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle detection systems face challenges in accurately distinguishing traffic lights from other illuminated objects in a traffic scene, leading to potential safety issues such as misinterpretation of brake lights as traffic lights, which can result in rear-end collisions.
Innovation Solution
A hierarchical ontology and machine learning model are used to detect and differentiate between traffic lights and other illuminated objects by computing probabilities based on categories within a hierarchical structure, ensuring accurate identification and state detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a detection system uses simple object identification methods, then the system complexity is low, but the accuracy of distinguishing traffic lights from other illuminated objects deteriorates
Solution Approach 1:
The patent segments the detection process into multiple hierarchical levels: first identifying illuminated objects, then classifying them into categories (traffic lights, brake lights, etc.), and finally determining specific states. This multi-stage segmentation allows the system to maintain manageable complexity while achieving high identification accuracy through progressive refinement at each level.
Solution Approach 2:
The patent introduces a hierarchical classification dimension that transforms the simple binary classification problem into a multi-dimensional categorization framework. By adding layers of classification (object type, specific category, state), the system achieves higher precision without proportionally increasing overall complexity, as each dimension builds upon the previous one.
2Productivity
If a detection system uses basic classification methods, then the processing speed is high, but the reliability of distinguishing traffic lights from brake lights deteriorates
Solution Approach 1:
The patent performs preliminary classification of illuminated objects into broad categories before detailed analysis. By first identifying objects as potential traffic lights, brake lights, or other light sources, and then applying more sophisticated discrimination only where needed, the system maintains high processing speed for obvious cases while achieving high reliability for ambiguous cases.
Solution Approach 2:
The patent applies different levels of analysis quality to different objects based on their characteristics. For clearly identifiable objects, simple classification suffices, maintaining high speed. For ambiguous objects resembling both traffic lights and brake lights, the system applies more rigorous multi-factor analysis, ensuring high reliability where it matters most without unnecessarily slowing down the entire system.
Data Source
AI summary
System, methods, and other embodiments described herein relate to accurately distinguishing a traffic light from other illuminated objects in the traffic scene and detecting states using hierarchical modeling. In one embodiment, a method includes detecting, using a machine learning (ML) model, two-dimensional (2D) coordinates of illuminated objects identified from a monocular image of a traffic scene for control adaptation by a control model. The method also includes assigning, using the ML model, computed probabilities to the illuminated objects for categories within a hierarchical ontology of environmental lights associated with the traffic scene, wherein one of the probabilities indicates existence of a traffic light instead of a brake light in the traffic scene. The method also includes executing a task by the control model for a vehicle according to the 2D coordinates and the computed probabilities.