Tire Tread Evaluation Using ML Contamination Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for electronically assessing the tread condition of vehicle tires are inadequate in distinguishing temporary contamination from reduced tread depth, leading to inaccurate assessments and potential premature tire replacement, especially in dirty or contaminated conditions common in commercial and work vehicle fleets.
Innovation Solution
A method using optical sensor data and machine learning-based identification to differentiate and ignore contaminated areas in the tread pattern, allowing for reliable and accurate evaluation of tread depth and condition, even in the presence of contaminants, through a combination of camera imaging and data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automated electronic evaluation devices are used to determine tread depth, then productivity and automation are improved, but measurement precision deteriorates due to contamination misinterpretation
Solution Approach 1:
The evaluation process is segmented into two independent analysis streams: one for tread depth measurement and another for contaminant detection. The contaminant detection module independently identifies contaminated areas, which are then excluded from the tread depth calculation. This segmentation allows automated evaluation to proceed while preventing contamination from degrading measurement precision.
Solution Approach 2:
A machine learning-based contaminant detection module acts as an intermediary between the optical sensor data and the tread depth evaluation. This intermediary analyzes the optical data to identify contaminants, separates contaminated regions from clean regions, and provides a filtered dataset for accurate tread depth measurement. The intermediary prevents contamination from directly affecting the measurement precision.
2Device complexity
If simple ruler-based tests are used to check tread depth, then device complexity is reduced, but measurement precision and reliability are insufficient for accurate assessment
Solution Approach 1:
The patent replaces mechanical ruler-based measurement with an automated optical sensor system combined with machine learning analysis. The optical sensor captures images of the tread pattern, and the machine learning model automatically measures tread depth and identifies contaminants. This substitution maintains simplicity from the user perspective while dramatically improving measurement precision and reliability through automated digital analysis.
3Productivity
If contaminated tread areas are included in the assessment, then productivity is maintained, but measurement precision deteriorates due to false contamination interpretation
Solution Approach 1:
The system performs preliminary contaminant detection and identification before conducting the final tread depth evaluation. By pre-identifying contaminated areas in the optical data, the system can exclude these regions from the measurement calculation. This preliminary action maintains productivity by automating the entire process while ensuring measurement precision by preventing contamination from skewing results.
Solution Approach 2:
The system dynamically changes the evaluation parameters by adjusting which regions of the tread are included in the measurement based on contaminant detection. When contaminants are detected, the system modifies the measurement mask to exclude contaminated areas, thereby changing the effective parameters of the evaluation. This allows the system to maintain high productivity through automation while adapting to local conditions to preserve measurement precision.
Data Source
Figure 1
Figure 2a~2b
AI summary
The invention relates to a method for evaluating the profile condition of a tread profile of vehicle tires (10) using an electronic condition assessment device (12), comprising the method steps: a) aligning the optical sensor arrangement (14) with a profile section (22) of the tread profile of the vehicle tire (10), b) recording optical sensor data of the profile section (22) with the optical sensor arrangement (14), and c) evaluating the recorded optical sensor data with the electronic data processing device (18) to evaluate the profile condition of the vehicle tire in the profile section (22), wherein in method step b) in addition to or as part of the optical sensor data, a camera image of the profile section (22) is recorded with the camera (16), wherein the electronic data processing device (18) is configured toto identify profile areas (24) in the profile section (22) blocked by impurities in the camera image using a machine learning-based identification module (26), wherein the identified blocked profile areas (24) are not taken into account or are taken into account to a reduced extent when assessing the profile condition in process step c).