License Plate Recognition via Intermediate Feature Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional license plate number recognition systems have high error rates due to the accumulation of errors in multiple independent steps and strict requirements for input images, making them sensitive to weather and lighting conditions, and lacking verification mechanisms, resulting in the need for manual verification.
Innovation Solution
A license plate number recognition method using a pre-trained convolutional neural network that extracts features, includes verification features, and verifies the license plate number through intermediate convolution results, reducing errors by outputting results only when verification is passed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional multi-step independent processing is used for license plate recognition, then the recognition process can be implemented, but error rates accumulate across steps and reliability deteriorates
Solution Approach 1:
The patent merges multiple independent processing steps (image normalization, pre-processing, character segmentation, character recognition) into a unified deep learning model. The convolutional neural network performs feature extraction, segmentation, and recognition in an integrated manner, eliminating error accumulation between separate steps while maintaining comprehensive processing functionality.
Solution Approach 2:
The patent implements a verification mechanism that extracts intermediate convolution results during feature extraction and uses them to verify the final license plate recognition results. This feedback loop checks whether extracted features meet expected characteristics, allowing the system to identify and correct potential errors before final output, thereby improving reliability.
2Measurement precision
If strict requirements are imposed on input images regarding angles and definitions, then recognition precision can be maintained, but adaptability to various monitoring scenes deteriorates
Solution Approach 1:
The patent employs a deep learning model that automatically adapts to varying image parameters including different angles, resolutions, lighting conditions, and weather scenarios. The convolutional neural network learns optimal feature representations across diverse parameter variations during training, enabling high recognition precision without imposing strict input requirements while maintaining broad scene adaptability.
3Productivity
If conventional recognition methods are used without verification mechanisms, then processing speed is maintained, but output error rate increases requiring manual verification
Solution Approach 1:
The patent implements a verification mechanism that extracts intermediate convolution results during feature extraction and uses them to verify the final license plate recognition results. This feedback loop checks whether extracted features meet expected characteristics, allowing the system to identify and correct potential errors before final output, thereby improving reliability.
Solution Approach 2:
The patent performs verification checks on intermediate convolution results before final recognition output is generated. By conducting preliminary verification on extracted features against expected characteristics, the system prevents erroneous results from being output, reducing the need for manual verification while maintaining processing efficiency.
Data Source
AI summary
A license plate number recognition method includes: extracting license plate number features of an image to be recognized including a license plate number, through a pre-trained convolutional neural network; extracting an intermediate convolution result during extracting the license plate number features, and extracting a first verification feature and/or a second verification feature according to the intermediate convolution result; verifying whether the license plate number features are correct according to the first and/or second verification features; if correct, outputting a predicted license plate number result according to the license plate number features. During the feature extraction process of the license plate number features, an intermediate feature is extracted as a verification feature to verify whether the extracted license plate number features are correct, and only when the verification is passed, outputting the license plate number result, which reduces the output error rate of the license plate number recognition result.