X-ray Test Image Generation via Virtual Hazard Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for training and testing operators of X-ray inspection facilities face delays and administrative burdens due to the need for realistic and regularly updated test X-ray images, which can be unrealistic and recognizable as such, and require additional inspection efforts when using fictitious or centrally generated images.
Innovation Solution
A device and method for generating test X-ray images based on actual inspection objects checked on-site, using an inspection object database and target item database to project virtual images of target items into X-ray images of inspected objects, ensuring the images are realistic and not recognizable as test images, and updating the database to maintain freshness and relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fictitious threat images (FTI) are used for training operators, then training can be conducted with standardized test images, but additional inspection efforts are required and delays occur during operations
Solution Approach 1:
The system performs preliminary actions by automatically generating test images from historical inspection data before training sessions. X-ray images of previously inspected luggage are retrieved and processed to embed hazardous items, creating ready-to-use test images that eliminate the need for manual image creation and prevent operational delays.
Solution Approach 2:
The system creates copies of real inspection images from the database and embeds hazardous items into these copies. This allows training images to be generated without affecting actual inspection operations, as the system works with replicated data rather than original inspection images.
2Adaptability or versatility
If centrally generated test images are used, then standardization is achieved, but operators may become familiar with test images reducing training effectiveness
Solution Approach 1:
The system implements dynamics by automatically updating the training image database with new inspection images. Test images are generated on-demand from the database, ensuring continuous variation and preventing operators from memorizing static test sets. The system adapts to maintain training effectiveness through automated image rotation and hazardous item embedding in diverse luggage contexts.
Solution Approach 2:
The system incorporates feedback mechanisms where inspection results from actual operations feed back into the training image database. Successfully inspected images are retrieved and converted into test images, ensuring that training materials reflect current operational patterns and hazards, thereby maintaining training relevance and effectiveness.
3Manufacturing precision
If test images are generated from actual inspection objects, then realism is improved, but additional inspection efforts and delays occur
Solution Approach 1:
The system performs preliminary actions by pre-processing and storing inspection images in a database during normal operations. When training images are needed, the system retrieves pre-stored images and embeds hazardous items computationally, avoiding time-consuming manual image creation and preventing operational delays.
Solution Approach 2:
The system replaces manual mechanical processes with automated computational methods. Instead of physically creating test images or manually inspecting objects for training purposes, the system uses automated image retrieval, digital hazardous item embedding, and database management to generate realistic test images efficiently.
4Adaptability or versatility
If a large number of centrally stored test images are maintained, then training variety is ensured, but administrative effort and costs increase
Solution Approach 1:
The system creates test images by copying and processing existing inspection images from the database rather than maintaining a separate large library of pre-created test images. This reduces administrative overhead while ensuring variety, as the system can generate numerous test images from a relatively small set of source inspection images by embedding different hazardous items and applying various transformations.
Solution Approach 2:
The system implements self-service by automatically generating test images on-demand from the inspection database without requiring manual administrative intervention. The automated processes retrieve images, embed hazardous items, and manage the test image pool, eliminating the need for dedicated administrative staff to curate and update training image libraries.
Data Source
AI summary
The present disclosure relates to a device and method for generating test X-ray images for nondestructively inspecting an inspection object, such as a piece of luggage or a package. The generation device includes: an inspection object database for storing X-ray image data of inspection objects generated using at least one X-ray inspection facility; a target item database containing image data of target items; and an image projection unit for projecting a virtual image of a target item from the target item database into an X-ray image from the inspection object database to generate a test X-ray image. The image projection unit is configured for generating test images based on an X-ray image of an inspection object, the X-ray image having been recorded prior to a first time period, and the associated inspection object having already been classified as safe and/or having been classified as safe during an automatic inspection.