Vehicle License Plate Recognition via Dual-Path Character Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vehicle license plate recognition methods often result in misreading due to incorrect character separation, leading to higher processing costs and reduced throughput, as they fail to effectively utilize structural characteristics of license plates.
Innovation Solution
A system and method that employs a combination of type-specific and universal character recognition using different classification criteria to accurately recognize vehicle license plate information, incorporating a license plate detecting module, character separation module, and character recognition module to generate and verify the license plate character string, and reject incorrect results based on comparisons between type-specific and universal recognition outcomes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional character separation and recognition methods are used, then the processing speed is maintained, but the recognition accuracy deteriorates due to misreading from incorrect character separation
Solution Approach 1:
The patent divides the character recognition process into two independent segments: type-specific character recognition and universal character recognition. Each segment uses different classification criteria and recognition methods. This segmentation allows the system to maintain high accuracy through multiple recognition paths while preserving processing speed by operating in parallel rather than sequential verification.
Solution Approach 2:
The patent changes the recognition parameters by employing two different classification criteria: one specific to license plate types and another universal criterion. This parameter diversification enables the system to cross-validate recognition results, improving accuracy without requiring repeated processing passes that would slow down throughput.
2Reliability
If multiple recognition methods are employed to improve accuracy, then the recognition reliability is improved, but the system complexity increases
Solution Approach 1:
The patent implements a universal character recognition module that can handle different license plate types using a common classification criterion. This universal module works alongside type-specific recognizers, providing multi-functionality that improves reliability through cross-validation while avoiding the need for completely separate recognition systems for each license plate type, thus controlling complexity.
Solution Approach 2:
The system employs feedback mechanisms where the results from type-specific character recognition and universal character recognition are compared and validated against each other. This mutual verification provides feedback that enhances reliability by identifying and correcting misreadings, while the feedback loop is designed to be efficient, preventing excessive complexity in the validation process.
3Measurement precision
If character separation is performed to extract individual characters, then the character recognition can be achieved, but misreading occurs when license plate type determination is incorrect
Solution Approach 1:
The patent segments the recognition process into type-specific and universal pathways, where each pathway independently processes character images. By comparing results from both pathways, the system can detect inconsistencies caused by incorrect license plate type determination and correct misreadings, thereby maintaining high accuracy while reducing the misreading rate.
Solution Approach 2:
The universal character recognition module acts as an intermediary that validates the results from type-specific recognition. When discrepancies are detected, the universal module provides an alternative interpretation, serving as a mediator that resolves conflicts and corrects misreadings without requiring complete reprocessing of the character separation stage.
Data Source
AI summary
Provided is a technology for recognizing vehicle license plate information which includes detecting a position of a license plate from a vehicle image obtained by imaging a vehicle, extracting individual character images by separating a plurality of characters from a region of the detected license plate in the vehicle image and extracting a license plate type of the license plate according to a predetermined criterion, recognizing a license plate character string on the basis of the extracted individual character images and the license plate type and outputting the recognized license plate character string.