X-Ray Visual Layer Separation for Obscured Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional X-ray imaging systems struggle to provide clear images of hidden objects due to superimposition with clutter, leading to operator fatigue and difficulty in accurately identifying materials and anomalies, especially in large or cluttered cargo containers and parcels.
Innovation Solution
A method and system for visual layer separation in X-ray images, allowing operators to select and modify areas of interest, applying algorithms to strip away clutter layers based on pixel uniformity and material characteristics, and correcting for beam hardening to reveal obscured objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional X-ray imaging is used to inspect cargo and baggage, then the inspection throughput can be maintained, but the image clarity and object detectability deteriorate due to superimposition of clutter
Solution Approach 1:
The patent segments the superimposed X-ray image into multiple layers representing different depth planes. By dividing the image into foreground, midground, and background layers, the system can separate target objects from clutter that would otherwise be superimposed together, improving both image clarity and object detectability without reducing inspection throughput.
Solution Approach 2:
The patent introduces a depth dimension to the traditional 2D X-ray image by creating multi-layered representations. This dimensional transformation allows operators to view objects at different depths separately, eliminating the information loss caused by superimposition while maintaining the speed of conventional X-ray inspection.
2Measurement precision
If operators manually examine each X-ray image to identify hidden objects, then detection accuracy can be maintained, but operator fatigue and processing time increase
Solution Approach 1:
The patent performs preliminary processing of X-ray images by automatically separating them into multiple depth layers before presentation to the operator. This preliminary action reduces the cognitive load and time required for manual examination, as operators can directly examine separated layers rather than searching through superimposed images, thereby maintaining detection accuracy while reducing fatigue and processing time.
Solution Approach 2:
The patent introduces an intermediate processing layer that automatically performs image segmentation and layer separation. This intermediary system handles the complex task of separating superimposed objects, allowing operators to focus on detection rather than image interpretation, thus maintaining accuracy while reducing time loss.
3Productivity
If automated systems are used to inspect high-volume cargo, then productivity increases, but false alarm rates increase due to difficulty in distinguishing threats from clutter
Solution Approach 1:
The patent segments images into multiple layers, allowing automated systems to analyze each layer separately for threat detection. This segmentation reduces false alarms by eliminating confusion from superimposed clutter, while maintaining high inspection throughput through automated processing of the separated layers.
Solution Approach 2:
The patent applies different analysis criteria to different layers of the segmented image. By tailoring the detection algorithm to each specific layer's characteristics rather than applying a uniform analysis to the entire superimposed image, the system reduces false alarms while maintaining productivity.
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
Enhances the clarity of X-ray images by separating target objects from clutter, enabling accurate material identification and reducing false alarms, thus improving detection efficiency and reducing operator fatigue.
Implementation Method 1
at least one X-ray scan image obtained from a pre-stored data
Implementation Method 2
correcting for beam hardening to reveal obscured objects
Data Source
AI summary
The present specification relates to a method for enabling an operator to perform visual layer separation, the method including: retrieving at least one X-ray scan image from a memory in data communication with an inspection system, wherein the image comprises a first area of pixels representative of a target object obscured by a clutter object and a second area of pixels representative of the clutter object; receiving a selection of the pixels representative of the first area; receiving a selection of the pixels representative of the second area; determining if the selected second area meets a predefined quality threshold; if the selected second area meets the predefined quality threshold, generating a modified at least one X-ray image; and if the selected second area does not meet the predefined quality threshold, prompting the operator to select a different second area of pixels representative of the clutter object.