Multi-Field Vehicle License Plate Recognition via Sensor Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing license plate recognition systems face challenges in acquiring and processing high-quality images of moving vehicles, especially in restricted locations like vehicle inspection lanes, due to limited field of view and varying ambient lighting conditions, which affects the accuracy and reliability of license plate identification.
Innovation Solution
A vehicle service system with multiple imaging sensors positioned to capture images from different fields of view, including leading and trailing surfaces, and oriented to accommodate varying viewpoints, along with a processing system that evaluates images in parallel to determine confidence levels and extract license plate information, prioritizing images based on quality and lighting conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single camera with fixed field of view is used, then the system structure is simple, but the license plate image quality is insufficient under varying lighting and vehicle positions
Solution Approach 1:
The system divides the single camera function into multiple cameras with different fields of view (first camera for approaching vehicles, second camera for departing vehicles). Each camera captures images from optimized angles for specific vehicle positions, ensuring high-quality license plate images regardless of vehicle movement stage.
Solution Approach 2:
The system adds a spatial dimension by positioning cameras at different locations and orientations. The first camera is positioned to capture approaching vehicles from one angle, while the second camera captures departing vehicles from another angle, creating multiple viewing dimensions to overcome lighting and position constraints.
2Reliability
If multiple cameras with different fields of view are deployed, then the license plate recognition accuracy is improved, but the number of images requiring processing increases
Solution Approach 1:
The system performs preliminary evaluation of captured images using confidence score calculation before full processing. The processor quickly assesses whether each image meets quality thresholds for license plate recognition, eliminating the need to process low-quality images in detail and reducing overall processing time.
Solution Approach 2:
The system applies different processing strategies to different images based on their quality characteristics. High-confidence images from optimal fields of view undergo full license plate recognition processing, while low-confidence images are either enhanced selectively or discarded, optimizing resource allocation.
3Reliability
If images are captured from multiple viewpoints, then the chances of acquiring usable license plate images are improved, but the device complexity and cost increase
Solution Approach 1:
The imaging system serves multiple functions: the first camera captures approaching vehicles for license plate recognition, the second camera captures departing vehicles, and together they provide comprehensive vehicle documentation. This multi-functionality justifies the increased sensor count by maximizing utility from each device.
Solution Approach 2:
The system dynamically selects which camera images to process based on real-time vehicle position and lighting conditions. The processor evaluates confidence scores from both cameras and selectively processes only the most promising images, adapting the processing workload to actual acquisition success rather than treating all images equally.
Data Source
AI summary
A vehicle service system incorporating a set of imaging sensors disposed in an inspection lane through which a vehicle is driven. A processor is configured with software instructions to capture a set of images from the set of imaging sensors and to evaluate the captured images according to a set of rules to identify images in which a license plate is visible on an observed surface of the vehicle. The processor is further configured with software instruction to extract license plate information from the identified images, assign a figure of merit to the extracted information, and generate an output in response to the assigned figures of merit.