Substrate Inspection Device Using Reference Image Difference Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing substrate inspection methods fail to accurately detect defects, as they rely on brightness ranges of reflected light, leading to false negatives when defects have brightness within predetermined allowable ranges.
Innovation Solution
An inspection device that calculates differences in gradation values between image data of a defect-free substrate and the substrate to be inspected, determining defects based on whether these differences fall within predetermined allowable ranges, and includes features like smoothing and moire removal to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If brightness range inspection is used to detect defects, then the inspection process is simple, but defects with brightness within allowable ranges cannot be detected
Solution Approach 1:
The system performs preliminary actions by acquiring image data of a substrate with no defect in appearance before inspecting the substrate to be inspected. This reference image data is stored and used for subsequent difference calculations, enabling the system to detect defects that would be invisible using traditional brightness range inspection alone.
Solution Approach 2:
The invention transitions from single-dimension brightness threshold inspection to a two-dimensional comparison approach. By calculating differences between corresponding pixels in reference image data and inspection image data, the system adds a temporal/dimensional dimension to the inspection process, enabling detection of subtle defects through change detection rather than absolute brightness evaluation.
2Measurement precision
If difference calculation between image data is performed, then defect detection accuracy is improved, but false positives from noise and moire patterns increase
Solution Approach 1:
The system converts harmful noise and moire patterns into detectable features by comparing reference image data with inspection image data. Since noise and moire patterns remain consistent between the reference and inspection images, their differences are minimized, while actual defects produce significant differences. This transforms the harmful effect of noise into a beneficial reference point for distinguishing real defects from artifacts.
Solution Approach 2:
The system implements feedback by using the reference image data (acquired from a known good substrate) as a baseline for comparison. The difference calculation provides feedback about deviations from the normal state, enabling the determination unit to distinguish between acceptable variations (noise, moire) and actual defects based on whether the differences exceed predetermined thresholds.
3Measurement precision
If image data of defect-free substrate is acquired for comparison, then detection accuracy for subtle defects is improved, but inspection time and data processing increase
Solution Approach 1:
The reference image data of a substrate with no defect in appearance is acquired in advance and stored for repeated use. This preliminary action eliminates the need to capture and process reference images during each inspection cycle, significantly reducing inspection time while maintaining high detection accuracy through consistent comparison against the stored reference data.
Data Source
AI summary
Surface image data of a non-defective sample substrate is acquired, and surface image data of a substrate to be inspected is acquired. Differences between gradation values are calculated for pixels of the surface image data of the substrate to be inspected and corresponding pixels of the surface image data of the sample substrate. A constant value is added to the difference between gradation values of each pixel. In the case where the value acquired by addition is in a predetermined allowable range, it is determined that there is no defect for the substrate to be inspected. In the case where the value acquired by addition is outside of the allowable range, it is determined that the substrate to be inspected is defective. A defect in appearance on the substrate to be inspected is detected based on a pixel of which the value is outside of the allowable range.