Tyre Defect Detection Using Edge Dilation to Filter False Alarms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current quality control methods for tyres in production processes are inefficient due to excessive computational requirements, leading to prolonged inspection times and false defect alarms caused by intrinsic variabilities in tyre production, such as mould variations and surface imperfections.
Innovation Solution
A method involving optical acquisition and processing of digital images, where edges from a sample tyre are compared to dilated edges of a reference tyre to filter out false alarms, using edge extraction, dilation, and restoration techniques to enhance defect detection reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If comparison-based defect detection is used on tyre surfaces, then defect identification capability is improved, but false alarms increase due to intrinsic variabilities like mould variations and production imperfections
Solution Approach 1:
The patent applies local quality by analyzing local edge characteristics and patterns rather than treating the entire tyre surface uniformly. The system identifies defect-specific edge features (such as irregular cuts or protrusions) distinct from normal production variations like mould joint lines or air outlet reliefs, enabling differentiated assessment of different surface regions based on their defect likelihood
Solution Approach 2:
The patent implements preliminary action by performing edge extraction and pattern recognition before final defect classification. The system pre-processes the tyre surface image to extract edge features, compares them against known defect patterns, and pre-identifies potential defect regions, thereby reducing false alarms during the final detection phase by eliminating non-defect variations in advance
2Measurement precision
If traditional image processing algorithms are used for defect detection, then defect identification capability is improved, but computational requirements and processing time increase excessively
Solution Approach 1:
The patent applies taking out by extracting only the critical edge features from the tyre surface images rather than processing the entire image data. The system isolates edge patterns that are indicative of defects and compares only these extracted features against reference patterns, significantly reducing computational requirements while maintaining defect detection accuracy
Solution Approach 2:
The patent implements segmentation by dividing the tyre surface analysis into distinct processing stages: edge extraction, pattern comparison, and defect classification. This segmented approach processes only relevant features at each stage rather than analyzing all image data comprehensively, thereby improving inspection speed without sacrificing detection capability
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
A method and apparatus for detecting defects on tyres in a tyre production process are described. The method comprises receiving (202) a reference image (lref) of a surface portion of a reference tyre substantially defects-free, comprises providing (203) a sample tyre to be analysed, comprises illuminating (204) a surface portion of the sample tyre with a raking light source, comprises acquiring (205) a sample image (lcmp) of the illuminated surface portion of the sample tyre, comprises extracting (207) the edges from the reference image and from the sample image and respectively generating an edge reference image (lref_edg) comprising the edges included in the reference image and an edge sample image (lcmp_edg) comprising the edges included in the sample image, comprises carrying out a dilation (208) of the edges of the edge reference image (lref_edg) and generating therefrom a dilated edge reference image (lref_edg_di), comprises comparing (209) the edge sample image (lcmp_edg) with respect to the dilated edge reference image (lref_edg_di) and generating an edge image of possible defects (Iedg_pd) comprising the edges that are included in the edge sample image and are not included in the dilated edge reference image, and comprises identifying as possible defects the edges included in the edge image of possible defects.