Defect Identification Using Luminance Normalization and Shape Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing surface defect identifying devices require multiple image capturing devices, leading to increased complexity and time consumption for defect inspection, and may erroneously detect non-defects due to luminance fluctuations across the image capturing range.
Innovation Solution
A defect identifying method and device that uses a single image capturing unit and pattern illumination to accurately identify defects by normalizing luminance, correcting luminance contrast based on shape and position variables, and comparing it to threshold values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple image capturing devices are used to capture images for defect inspection, then measurement precision is improved, but device complexity increases and inspection time is extended
Solution Approach 1:
The patent divides the inspection process into two stages: first capturing a full-range image to identify potential defect regions, then capturing detailed images only of those specific regions. This segmentation approach maintains high detection accuracy while reducing the need for multiple simultaneous image capturing devices throughout the entire inspection area.
Solution Approach 2:
The patent performs preliminary defect region identification using a full-range image before conducting detailed inspection. This preliminary action allows the system to focus subsequent detailed imaging only on areas where defects are suspected, thereby reducing overall system complexity and inspection time while maintaining measurement precision.
2Measurement precision
If multiple image capturing devices are used to capture images for defect inspection, then measurement precision is improved, but inspection time is extended
Solution Approach 1:
The inspection process is segmented into a quick full-range scan followed by targeted detailed imaging only of suspected defect regions. This reduces total inspection time compared to capturing complete detailed images of the entire surface, while maintaining measurement precision through focused detailed examination of actual defect areas.
Solution Approach 2:
A preliminary full-range image capture identifies potential defect locations before detailed inspection begins. This preliminary action eliminates the need to spend time capturing and processing detailed images of entire defect-free areas, thereby reducing inspection time while preserving measurement precision for actual defects.
3Productivity
If binarization with a preset threshold value is used for defect detection, then productivity is improved, but measurement precision deteriorates due to erroneous detection
Solution Approach 1:
The patent applies different processing approaches to different regions of the image: full-range images use threshold-based binarization for efficient overview, while detailed images of specific defect regions use more sophisticated analysis methods. This local quality differentiation maintains productivity through efficient processing of most areas while improving measurement precision for critical defect regions.
Solution Approach 2:
The inspection process is segmented into two processing stages: threshold-based binarization for the full-range overview image to quickly identify potential defects, and more precise analysis methods for detailed images of specific regions. This segmentation allows the system to maintain high productivity through efficient threshold processing while achieving high measurement precision through detailed analysis of suspected defect areas.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A defect identifying method includes: an image capturing process of acquiring a captured image of a measurement object; a defect candidate detecting process of detecting a defect candidate; a shape information acquiring process of acquiring shape information of a position of the defect candidate; a normalizing process of normalizing luminance of the captured image; an identifying value acquiring process of acquiring an identifying value on the basis of the normalized luminance; a correction value acquiring process of acquiring a correction value corresponding to the shape information; a correcting process of correcting either the identifying value or a threshold value with the correction value; and an identifying process of identifying whether or not the defect candidate is a defect.