Dynamic Digital Mask for Container Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for inspecting translucent or transparent containers with visual patterns struggle to detect defects accurately due to random orientation and varying contrast, leading to reduced sensitivity and failure to analyze patterns properly.
Innovation Solution
An optical method that illuminates containers and produces images with determined pixel intensity levels, applies geometric transformations to digital masks based on pattern position and orientation, and updates the masks to ensure coincidence with processing zones for effective defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If digital masks are used to inspect containers with visual patterns, then defect detection is performed, but sensitivity is reduced in pattern areas due to random orientation and contrast variability
Solution Approach 1:
The digital mask is made dynamic by automatically adapting its parameters (position, orientation, contrast thresholds) based on the detected orientation and characteristics of visual patterns in each container image. This allows the inspection system to adjust to varying pattern orientations rather than using a fixed mask, thereby maintaining high defect detection sensitivity across all pattern configurations.
Solution Approach 2:
The system changes key parameters of the digital mask including orientation angle, position coordinates, and contrast thresholds based on the detected visual pattern characteristics. By dynamically adjusting these parameters to match each container's specific pattern orientation, the system overcomes the sensitivity reduction problem caused by random pattern orientations.
2Reliability
If image processing is applied outside visual patterns to avoid considering patterns as defects, then false positives are reduced, but patterns containing actual defects are not analyzed
Solution Approach 1:
The inspection system applies different processing strategies to different regions: within visual patterns, it uses adapted digital masks that account for pattern characteristics, while in non-pattern areas, it uses standard processing. This local differentiation allows the system to analyze patterns for defects without generating false positives from normal pattern variations.
Solution Approach 2:
The system dynamically determines the appropriate processing method for each region based on whether it contains a visual pattern. By detecting pattern boundaries and characteristics, the system adaptively switches between pattern-specific processing and standard processing, ensuring that actual defects within patterns are detected while normal pattern features are not misidentified as defects.
3Adaptability or versatility
If multiple digital masks are used to account for pattern particularities at different rotation positions, then all pattern variations are covered, but system complexity increases
Solution Approach 1:
Instead of using multiple static digital masks for different rotation positions, the system employs a single dynamic digital mask that automatically adjusts its parameters based on the detected pattern orientation in each image. This dynamic approach provides comprehensive coverage of all pattern variations while maintaining low system complexity, as only one adaptable mask is needed rather than multiple fixed masks.
Solution Approach 2:
The system changes the orientation parameter and position parameters of the digital mask based on detected pattern characteristics, eliminating the need to pre-create multiple masks for different rotations. This parameter adaptation approach achieves full pattern variation coverage with a single mask, significantly reducing system complexity.
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 method enables reliable detection of defects within visual patterns regardless of their position, improving sensitivity and accuracy by adapting to the orientation and morphology of the patterns, thus overcoming the limitations of previous techniques.
Implementation Method 1
The camera produces an image thanks to the light passing through the containers. This principle of lighting is said by transmission.
Data Source
Figure 1~2
Figure 2A~3
AI summary
The invention relates to an optical method for inspecting containers (2), said method consisting in: producing at least one image (10) of each container; determining, in the image of the container, at least one search area (Zr) wherein at least one visual motif appears (3); producing a digital mask (Mi) for at least one area (Zt) for processing the images comprising at least the visual motif (3); and comparing at least each pixel of the image processing area with a digital mask (Mi). According to the invention, the method consists in: selecting at least one visual motif (3) pertaining to the container; determining the position and the orientation of the visual motif (3) selected in said search area (Zr) of the image of the container; applying a geometric transformation (T) to the digital mask (Mi) or the processing area (Zt) in such a way as to be able, during the processing step, to place the mask (Mi) and the processing area (Zt) in a position of coincidence; and applying an image processing operation to each pixel of the processing area (Zt), depending on the value of intensity of the coinciding pixel of the digital mask (Mi).