Vehicle Classification Using Motion Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicle classification systems based on laser scanners and in-ground sensors are expensive and not easily scalable for monitoring streets and bridges, while automated enforcement requires a low-cost and computationally efficient method to distinguish trucks or buses from other vehicles for traffic law enforcement.
Innovation Solution
A computer-implemented method and image processing system that generates and associates motion vectors with vehicle clusters to classify vehicles, allowing for low-cost and efficient video-based classification, which can be integrated into existing video cameras for real-time enforcement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional laser scanner and in-ground sensor systems are used for vehicle classification, then measurement precision and reliability are improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent replaces mechanical sensor systems (laser scanners and in-ground sensors) with a video-based optical system. The image capturing device records video of vehicles, and motion vectors are extracted from the video frames to classify vehicles. This substitution reduces device complexity and cost while maintaining sufficient classification accuracy for traffic enforcement applications.
2Measurement precision
If conventional laser scanner and in-ground sensor systems are used for vehicle classification, then measurement precision is improved, but ease of manufacture and scalability worsen
Solution Approach 1:
The patent replaces expensive, difficult-to-deploy mechanical sensor systems with video cameras and image processing software. This substitution dramatically improves ease of manufacture and scalability, allowing the system to be easily deployed across multiple locations for city-wide traffic monitoring and enforcement.
3Ease of operation
If video-based classification is used instead of conventional systems, then ease of operation and scalability are improved, but measurement precision may worsen
Solution Approach 1:
The patent changes the measurement parameters from detailed physical measurements (axle number, height, width, weight, length, profile, volume) to motion-based parameters (motion vectors from video frames). This parameter change enables easier operation and scalability while providing sufficient precision for distinguishing trucks and buses from other vehicles for traffic law enforcement purposes.
4Measurement precision
If detailed vehicle information is captured using conventional systems, then measurement precision is improved, but loss of time and computational requirements worsen
Solution Approach 1:
The patent extracts only the essential information needed for traffic enforcement (vehicle type classification) from video motion data, rather than capturing all detailed vehicle parameters. This extraction approach reduces processing time and computational requirements while providing sufficient accuracy for determining traffic violations.
Data Source
AI summary
This disclosure provides methods and systems of classifying a vehicle using motion vectors associated with captured images including a vehicle. According to an exemplary method, a cluster of motion vectors representative of a vehicle within a target region is analyzed to determine geometric attributes of the cluster and/or measure a length of a detected vehicle, which provides a basis for classifying the detected vehicle.