Glassware Defect Detection Using Offset Backlighting Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting transparent defects in glass articles, such as small bubbles and inverted folds, are inadequate due to low contrast and difficulty in detection, especially in articles with non-homogeneous thickness and complex geometries, which can lead to safety issues in industries like perfumery and pharmaceuticals.
Innovation Solution
A method involving backlighting with a pair of spatially offset lighting patterns, where images from each pattern are combined to enhance contrast and detect defects, using a single light source that generates patterns of different colors and adjusting the ratio of strip width to pattern pitch for optimal defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single light source with structured pattern is used, then device complexity is reduced, but measurement precision deteriorates due to inability to detect defects in dark zones
Solution Approach 1:
The patent divides the inspection process into multiple sequential stages: first inspection with structured light pattern, then second inspection with homogeneous light pattern. This segmentation allows each lighting condition to specialize in detecting different defect types, resolving the contradiction by maintaining simple device structure while achieving comprehensive defect detection through temporal rather than spatial division.
Solution Approach 2:
The patent employs periodic alternation between two lighting patterns (structured and homogeneous) to illuminate the article sequentially. This periodic action enables the system to capture images under different lighting conditions, allowing detection of both structured-pattern-visible defects and homogeneous-pattern-visible defects, thereby maintaining measurement precision without requiring permanently complex lighting hardware.
2Measurement precision
If structured light pattern is used, then contrast of certain defects is improved, but defects in dark zones remain undetected
Solution Approach 1:
The inspection process is segmented into two distinct phases: first inspection using structured light pattern to enhance contrast of certain defects, and second inspection using homogeneous light pattern to illuminate dark zones. This segmentation ensures that no defect type is missed while maintaining the contrast-enhancing benefits of structured lighting for applicable defects.
Solution Approach 2:
The patent implements continuous inspection by seamlessly alternating between structured and homogeneous lighting patterns. The article moves continuously through the inspection zone while receiving sequential illumination from both lighting types, ensuring uninterrupted and comprehensive defect detection without losing information about any region or defect type.
3Measurement precision
If structured light pattern is used, then small defects can be detected, but articles with non-constant thickness produce difficult-to-analyze images
Solution Approach 1:
The patent segments the detection task into two specialized inspections: structured light inspection optimized for detecting small defects through contrast enhancement, and homogeneous light inspection that produces simpler images easier to analyze for articles with thickness variations. This segmentation allows each method to excel at its specialized function without suffering from the other's limitations.
Solution Approach 2:
The patent changes the lighting parameter (from structured pattern to homogeneous pattern) depending on the inspection stage and article characteristics. By dynamically adjusting the lighting parameter, the system optimizes for small defect detection when using structured patterns, and simplifies image analysis when using homogeneous patterns, thereby reducing overall analysis difficulty while maintaining small defect detection capability.
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 improves the detection of transparent defects by increasing contrast and allowing for precise identification of flaws, even in complex geometries, with a high success rate and compatibility with current production rates.
Implementation Method 1
backlighting the article in question by means of one or more light sources
Implementation Method 2
defects which appear weakly contrasted by backlighting
Data Source
Figure 1~2c
Figure 3A~4
Figure 5A~6
AI summary
Method for detecting so-called low-contrast defects in glassware articles (3), involving: back-lighting said article (3) by means of one or more light sources (6) according to at least one pair of spatially offset similar lighting patterns; capturing, using two cameras (17, 18), at least one pair of images of said backlit article according to each of the patterns of said at least one pair of patterns; combining the images of each of said pairs to form at least one composite image; detecting the zones having the strongest contrast within said composite images.