Skip to Main Content

ALERT has received numerous awards for supplemental research aligned with our mission and vision of a world protected from the catastrophic consequences of explosives-related threats. Through collaborations with government agencies, partner universities, and industry, ALERT researchers have undertaken and completed a range of additional directed research efforts, as detailed below.

Computed Tomography Segmentation Advanced Algorithm Research

During the annual ADSA workshops, participants agreed that the segmentation algorithms used in automated threat recognition (ATR) need improvement to yield more precise features of explosives. As a result, researchers started working in 2010 to address the artifacts in CT images, such as blurring, streaking, and low-frequency shading, and to develop better segmentation algorithms for aviation security. They used non-threat scans on medical CT scanners for the purpose. The initiative successfully engaged third-party contributors and achieved its objectives with minimal resources.

Computed Tomography Segmentation Advanced Algorithm Research Final Report

Research and Development of Reconstruction Advances in CT-Based Object Detection Systems Project

Researchers began work to develop advanced reconstruction algorithms for explosive detection systems based on computerized tomography in 2012. Researchers developed new algorithms for medical CT scans using iterative reconstruction, sinogram processing, dual-energy decomposition, and simultaneous reconstruction/segmentation. The algorithms improve detection performance by reducing streak artifacts and improving contrast between objects. The public domain database created to share results allows third parties to develop technologies that can be transitioned to deployed equipment.

Research and Development of Reconstruction Advances in CT-Based Object Detection Systems Project Final Report

Advances in Automatic Target Recognition For CT-Based Object Detection Systems

In 2012, five groups of researchers initiated work to develop automated target recognition (ATR) algorithms for CT-based explosive detection systems. The primary goal of the project was to achieve a probability of detection greater than 90% and a probability of false alarm less than 10%. The project focused on using non-classified datasets and requirements, encouraging third parties to develop technologies that can be transitioned to deployed equipment, thus improving the overall effectiveness of the explosive detection systems.

Advances in Automatic Target Recognition For CT-Based Object Detection Systems Final Report

Improved Millimeter Wave Radar Characterization of Concealed Low-Contrast Body-Bourne Threats

ALERT received funding in 2012 to conduct Advanced Imaging Technology (AIT) research. Led by Carey Rappaport and Jose Martinez-Lorenzo the contract aimed to adapt two algorithm projects for industrial partners. The projects included nearfield radar material characterization of concealed body-borne dielectric threats (such as explosives) and automatic body surface reconstruction using time-of-flight monostatic radar.

Maturation and Validation of Dielectric Characterization Algorithms

A continuation of the Improved Millimeter Wave Radar Characterization of Concealed Low-Contrast Body-Bourne Threats research, this project aimed to create algorithms for Advanced Imaging Technology (AIT) material characterization. The goal was to detect foreign objects attached to people and minimize the need for physical pat-downs at airport security checkpoints. The program investigated new approaches to process data generated by AIT systems to determine the thickness and dielectric constant of foreign objects. The resulting algorithms were then refined to support the development of a software-based tool that could be used with existing AIT systems in the field.

Trace Explosives Sampling Efficiency and Baseline Performance

Researchers set out to create standardized procedures and methods to measure the efficiency and performance of contact sampling for trace explosives detection in 2015. A direct result of the Trace Explosives Sampling for Security Applications (TESSA) workshop series hosted by ALERT and led by Professor Stephen Beaudoin of Purdue University, this work resulted in reference materials, procedures for using the reference materials, and a limited database of contact sampling.

TESSA01 Presentations

Comprehensive Database of Contact Explosives Sampling Efficiency and Baseline Performance

This project began in 2016 as an extension of previous Trace Explosives Sampling Efficiency and Baseline Performance research to create a common language and standardized method to measure the sampling efficiency of explosives trace detection devices (ETDs). The prior work resulted in reference materials, procedures for using the reference materials, and a limited database of contact sampling. Researchers expanded the existing database by increasing the number and type of sample traps and explosives used in the testing process, focused on creating commercial off-the-shelf traps and reagent-enhanced traps.

TESSA02 Presentations

Video Anomaly Sensing and Tracking

Researchers began work to identify and track potential risks in vulnerable venues, detect counter-flow within a specific area, and tag and track a person of interest across multiple cameras. The system was designed to identify an individual in a single camera's view and re-identify them in each sequential camera's view. The developed video analytics technology is now used by the Transportation Security Administration to monitor and intercept potential threats at airports.

Video Anomaly Sensing and Tracking Final Report

Research and Development of Algorithms for Improved Image Quality for Checkpoint Explosive Detection Systems

This research led to the development of adaptive automatic target recognition (AATR) algorithms that are easily configured to add new targets without retraining and retesting the ATR. The resulting algorithm can add or remove targets, vary the minimum target mass and sheet thickness, and balance the probability of detection and false alarm. Researchers presented the new AATR technology in 2018, where experts discussed the algorithms' applicability to certified explosives detection equipment and the steps needed for further research in this area.

Novel Features and Emerging Technologies for Opioid Detection

In 2019, ALERT initiated research to aid the opioid detection prize competition launched by the Department of Homeland Security (DHS) Science and Technology Directorate. Researchers aimed to identify new signatures and predict the optimal detection performance for opioids that enter the country through international mail. The results equipped DHS to determine which new technologies to develop into prototypes to demonstrate effective detection capability.

Simulated Air Cargo Skids

This research developed a mathematical description of simulated air cargo packed on skids. Researchers use these mathematical descriptions alongside existing X-ray simulation tools to create simulated projections of air cargo contents. The expected outcome is a software tool that can be used by the DHS Science and Technology Directorate and the Transportation Security Laboratory to improve the detection capabilities of current inspection systems and detect anomalies more effectively.