X-Ray Inspection Validation Using Pseudo-Defect Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing X-ray inspection systems require manual validation using test samples, which increases worker workload and risks allowing defective products to pass through during production, especially for foreign object detection and shape abnormalities.
Innovation Solution
An X-ray inspection apparatus generates pseudo-defect images by altering pixel values in inspection images to simulate defects, allowing automated validation without physical samples, and includes a control unit to manage sorting and conveyance based on these images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual validation using test samples is performed, then inspection validity can be checked, but worker workload increases and risk of defective products passing through increases
Solution Approach 1:
The patent creates a virtual copy of the test sample by generating a pseudo-defect image from the inspection image of a normal product. This virtual copy contains simulated defect information that replicates the characteristics of actual defective products, allowing validation without physical test samples. The control unit generates the pseudo-defect image by processing the inspection image data, creating a digital replica that can be used for validation purposes.
Solution Approach 2:
The patent replaces the mechanical process of physically handling and running test samples through the inspection system with a digital image processing system. Instead of manually preparing and conveying physical test articles, the control unit automatically generates pseudo-defect images from inspection image data, substituting mechanical operations with computational processes to achieve the same validation function.
2Reliability
If manual validation using test samples is performed, then inspection validity can be checked, but risk of defective products passing through increases
Solution Approach 1:
The patent creates a virtual copy of the test sample by generating a pseudo-defect image from the inspection image of a normal product. This virtual copy contains simulated defect information that replicates the characteristics of actual defective products, allowing validation without physical test samples. The control unit generates the pseudo-defect image by processing the inspection image data, creating a digital replica that can be used for validation purposes.
Solution Approach 2:
The patent performs validation checks in advance by generating pseudo-defect images before actual production runs. The control unit proactively creates virtual test cases that simulate potential defects, allowing the inspection system to be validated and calibrated before real products are processed, thereby preventing defective products from passing through undetected.
3Ease of operation
If automated validation using pseudo-defect images is implemented, then worker workload is reduced, but device complexity increases
Solution Approach 1:
The control unit is designed to perform multiple functions: it processes inspection images, generates pseudo-defect images for validation, and manages the overall inspection workflow. By making the control unit multi-functional, the patent avoids adding separate dedicated hardware for validation purposes, thereby reducing overall system complexity while still achieving automated validation.
Solution Approach 2:
The patent replaces the mechanical process of physically handling and running test samples through the inspection system with a digital image processing system. Instead of manually preparing and conveying physical test articles, the control unit automatically generates pseudo-defect images from inspection image data, substituting mechanical operations with computational processes to achieve the same validation function.
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
Reduces the risk of defective products passing through and minimizes worker workload by automating validation checks using pseudo-defect images, ensuring accurate detection of defects and appropriate product sorting.
Implementation Method 1
an irradiation unit that irradiates the article, which is conveyed by the conveying unit, with X-rays; a sensor that detects the X-rays that have transmitted through the article
Data Source
AI summary
An X-ray inspection apparatus includes a transport unit that transports an article; an irradiation unit that irradiates the article, which is transported by the transport unit, with X-rays; a sensor unit that detects the X-rays; and a control unit that generates an inspection image including the article from a detection result of the X-rays by the sensor unit, and that inspects whether the article includes an abnormality, based on the inspection image. The control unit generates a pseudo-defect image including a virtual abnormality by changing pixel values of some pixels constituting the inspection image generated when the article is inspected, and checks validity of the inspection based on the pseudo-defect image.