Mass Estimation via Anisotropic Erosion for CT Contraband Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting contraband in containers using computed tomography (CT) data often generate false alarms due to the lack of consideration for partial volume effects, anisotropic effects, and CT beam hardening, as well as undersegmentation issues.
Innovation Solution
A method and system that utilize anisotropic erosion and dilation operators, along with histogram calculations, to accurately estimate the mass of objects within containers by defining a perimeter and calculating a CT number, thereby reducing false alarms and improving segmentation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If known methods are used to detect concealed objects in CT data, then the detection process can be performed, but false alarms are generated due to lack of consideration for partial volume effects, anisotropic effects, and CT beam hardening
Solution Approach 1:
The patent applies parameter changes by correcting CT numbers for beam hardening effects and using corrected CT numbers in mass estimation calculations. The system adjusts the CT number parameter based on the object's position and material properties to compensate for beam hardening, thereby improving measurement precision while maintaining detection accuracy.
Solution Approach 2:
The patent implements local quality by applying different correction factors and processing methods to different regions of the CT data. The system identifies objects with different material properties and applies location-specific corrections for partial volume effects and beam hardening, improving both detection accuracy and mass estimation precision locally throughout the scanned volume.
2Productivity
If known methods are used to detect concealed objects, then the detection process can be completed, but undersegmentation occurs and false alarms are generated
Solution Approach 1:
The patent applies segmentation by dividing the CT data into distinct regions based on density thresholds and object characteristics. The system segments objects from the background and from each other using improved thresholding methods that account for partial volume effects, enabling accurate mass estimation while maintaining high detection throughput.
Solution Approach 2:
The patent implements preliminary action by performing pre-processing steps before mass estimation, including correction of beam hardening effects and adjustment of CT numbers. These preliminary corrections are applied to the entire dataset before segmentation and mass calculation, preventing undersegmentation and false alarms while maintaining efficient processing speed.
3Adaptability or versatility
If CT scanning is performed to detect contraband, then the detection capability is provided, but false alarms are generated due to anisotropic effects and partial volume effects
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting CT number correction factors based on object properties and scanning conditions. The system modifies the CT number parameter for each object based on its density, size, and position, thereby adapting the detection capability to different object types while reducing false alarms caused by anisotropic effects and partial volume effects.
Data Source
AI summary
A system and method for identifying an object based on its estimated mass. In one aspect, a method for estimating a mass of an object is provided. The method includes acquiring image data including a plurality of image elements, calculating a histogram based on the image data, calculating a computed tomography (CT) number of the object using an anisotropic erosion operator, and determining a perimeter of the object. The method also includes calculating an estimated mass of the object using the CT number and a first subset of image elements of the plurality of image elements, the first subset of image elements defined by the perimeter of the object, and outputting at least one of the estimated mass of the object and an image including the object.