Multiple and Multi Sensor Research for Explosive Hazard Detection
Abstract
The nature of the explosive hazard threat t is continuously evolving, and hence, the sensor s and mechanisms to detect these threats must also evolve. The research tea m at the University of Missouri has been involved in research and development of algorithmic approaches to detect a variety of ex plosive hazards in various environments utilizing an assortment of traditional and novel no n-downward looking sensors for over a dozen years. The targets of interest have changed size and shape, and the emplacement of these targets have moved from being buried to being placed road side with increasing amount of obscuration. Additionally, the sensors used to detect these objects and separate them from false alarms have also changed dramatically over this time. Some traditional sensing modalities, like infrared imagery, provide a baseline on which algorithms have been developed. In an effort to increase detection and lower false alarms, new sensors come into the picture and sometimes go for various reasons. To keep making progress from basic and applied research perspectives necessitates a flexible but grounded program. The overall Army effort involves personnel from U.S Army Communications-Electronics Research, Development and Engineering Center s Night Visio n and Electronic Sensors Directorate (CE RDEC NYESD) along with various academic and industrial groups. Algorithms must be developed that recognize and utilize the physics and phenomenology of target, environment, and sensor suite to maximize performance and further the overall mission of the entire R&D team. Our research squad has been ab le to adapt to these changing scenarios and is positioned to continue to make significant strides in the future. Hence, the goal of this proposal is to perform basic and applied research into multiple and multi sensor explosive hazard detection. Specifically, we propose to perform research into sensor-based explosive hazard detection, being adaptive to new sensor types and problem domains. This includes new feature extraction and classification approaches for traditional 2-D as well as 3-D imaging modalities. We will also investigate multiple information source fusion to increase detection and/or lower false alarm rate, and research methods to characterize the environment to adapt feature extraction and classification to the environment including road following and environment characterization based on novel sensor s (e.g., LID AR, 3- D Radar) or methodologies, such as using robot platforms, remote sensing, drones. Finally, as partners with NYESD, we will rigorously test all of our algorithms and approaches both in- house and through blind testing by the Army.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2018
- Source ID
- W911NF1710183
Entities
People
- James M. Keller
Organizations
- Army Contracting Command
- United States Army
- University of Missouri