Tyre Sidewall Marking Classification Under Low-Contrast Wear
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for automating the reading of embossed and engraved markings on vehicle tires are inefficient due to high hardware costs, calibration challenges, and the need for human assistance, especially in outdoor conditions with low contrast and wear, and require significant processing resources for high-resolution images.
Innovation Solution
The method involves generating feature maps using histogram equalization and adaptive histogram equalization to enhance image contrast, reducing the processing resources needed for training and classification, allowing for real-time classification on a CPU rather than a GPU, and using a Faster R-CNN with three image channels for improved accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional OCR techniques are used on unenhanced images, then hardware costs are reduced, but accuracy is insufficient due to low contrast and text wear
Solution Approach 1:
The patent applies histogram equalization and adaptive histogram equalization to enhance image contrast before OCR processing. This preliminary image enhancement action prepares the low-contrast, worn text images to have improved visibility and contrast, enabling accurate recognition without requiring expensive specialized hardware
Solution Approach 2:
The patent introduces histogram equalization as an intermediary processing step between image capture and OCR recognition. This intermediary technique transforms the original low-contrast images into enhanced images with improved text visibility, bridging the gap between conventional hardware capabilities and OCR accuracy requirements
2Measurement precision
If high-resolution images are used for accurate text detection, then measurement precision is improved, but processing time and computational resources increase
Solution Approach 1:
The patent performs histogram equalization on the entire image before detecting text regions. This preliminary contrast enhancement makes text regions more distinguishable, enabling faster and more accurate region proposal in subsequent Faster R-CNN processing, thereby reducing overall processing time while maintaining high detection accuracy
Solution Approach 2:
The patent uses Faster R-CNN to segment the enhanced image into candidate text regions before detailed classification. This segmentation approach focuses computational resources on relevant areas only, reducing processing time compared to analyzing the entire high-resolution image uniformly
3Productivity
If Faster R-CNN is trained on unenhanced images, then training data preparation is simpler, but training duration and computational resources are excessive
Solution Approach 1:
The patent enhances all training images using histogram equalization before feeding them to the Faster R-CNN model. This preliminary enhancement ensures that the model learns from high-contrast images with clear text features, improving learning efficiency and reducing training duration while maintaining consistent processing across the entire training dataset
4Measurement precision
If structured light systems are used for indoor inspection, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent replaces complex mechanical structured light systems with a simpler digital image processing approach. Instead of using physical structured light projection and complex optical setups, the patent uses standard imaging combined with histogram equalization to achieve enhanced text visibility and accurate recognition, significantly reducing system complexity and cost
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces the duration of training and processing resources required, achieving higher test accuracy and enabling real-time classification of tire sidewall markings on a CPU, while maintaining high accuracy even in challenging outdoor conditions.
Implementation Method 1
a first image channel is generated from the digital image data by performing histogram equalisation on a first portion of the digital image data
Implementation Method 2
a second image channel is generated by performing adaptive histogram equalisation on the first portion of the digital image data
Data Source
Figure 1~2
Figure 3
Figure 4~6
AI summary
A computer implemented method is proposed for classifying one or more embossed and/or engraved markings on a sidewall of a tyre into one or more classes comprising digital image data of the sidewall of the tyre. The method comprises generating a first image channel from a first portion of the digital image data relating to a corresponding first portion of the sidewall of the tyre. Generating the first image channel comprises performing histogram equalisation on the first portion of the digital image data to generate the first image channel. The method further comprises generating a first feature map using the first image channel and applying a first classifier to the first feature map to classify said embossed and/or engraved markings into one or more first classes.