License Plate Character Estimation via Multi-Image Consensus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for automated license plate recognition in law enforcement are inefficient and inaccurate due to poor results from object detection algorithms and optical character recognition errors, especially in the presence of camera noise and varying perspectives.
Innovation Solution
A system that captures multiple images of a license plate, determines candidate characters using character recognition techniques, and selects target characters based on cross-correlation and longest common subsequence methods to consolidate results and output a single accurate representation of the license plate characters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images are captured and processed using conventional OCR techniques, then the quantity of character recognition attempts increases, but the accuracy and efficiency of license plate recognition deteriorates due to noise and perspective variations
Solution Approach 1:
The patent segments the license plate recognition problem into distinct phases: candidate character detection from individual images, followed by grouping and consensus determination across multiple images. This segmentation allows each phase to be optimized independently, improving both accuracy and efficiency.
Solution Approach 2:
The patent merges results from multiple image processing operations by grouping candidate characters across images and applying consensus algorithms. This combining approach leverages information from all images to produce a more accurate final recognition while maintaining efficiency through structured processing.
2Productivity
If conventional OCR techniques are applied to each image independently, then the processing speed is maintained, but the reliability of character recognition deteriorates due to camera noise and perspective distortion
Solution Approach 1:
The patent performs preliminary candidate character detection on each image independently to maintain processing speed, then applies grouping and consensus determination as a preliminary consolidation step before final recognition. This preliminary action on individual images followed by collective processing improves reliability without sacrificing overall speed.
Solution Approach 2:
The patent implements a feedback mechanism where candidate characters from multiple images are compared and consolidated through grouping operations. The consensus determination process uses feedback from all candidate detections to refine the final character identification, improving reliability while maintaining efficiency through structured iterative processing.
3Loss of information
If candidate characters are determined from multiple images without consolidation, then the quantity of data increases providing more information, but the complexity of determining actual characters increases due to varying perspectives and noise
Solution Approach 1:
The patent segments the complex consolidation problem into manageable steps: first grouping candidate characters by position and appearance across images, then applying consensus determination within each group. This segmentation reduces processing complexity while retaining information from all images through systematic organization.
Solution Approach 2:
The patent changes the processing parameters by transitioning from individual image analysis to grouped consensus analysis. By organizing candidate characters into groups based on spatial and visual parameters, the system manages information from multiple images efficiently, reducing complexity while maintaining information retention through structured parameter-based grouping.
Data Source
AI summary
Implementations generally relate to estimating grouped observations. In some implementations, a method includes detecting a license plate of a vehicle. The method further includes capturing a plurality of images of the license plate. The method further includes determining candidate characters from the images of the license plate. The method further includes selecting target characters from the candidate characters based on one or more predetermined selection criteria.