Nuclear quadrupole resonance system for landmine detection in Antioquia
Colombia ranks second in the world by number of victims from landmines; and in Colombia, Antioquia is the most affected department. As most landmines in Antioquia do not have metallic parts, metal detectors became useless, however, in most of those mines, the explosive mixture includes ammonium nitr...
- Autores:
-
Cardona Rendon, Lorena
- Tipo de recurso:
- Doctoral thesis
- Fecha de publicación:
- 2017
- Institución:
- Universidad Nacional de Colombia
- Repositorio:
- Universidad Nacional de Colombia
- Idioma:
- spa
- OAI Identifier:
- oai:repositorio.unal.edu.co:unal/59100
- Acceso en línea:
- https://repositorio.unal.edu.co/handle/unal/59100
http://bdigital.unal.edu.co/56322/
- Palabra clave:
- 62 Ingeniería y operaciones afines / Engineering
nuclear quadrupole resonance
resonancia nuclear en cuadrupolo
NQR
landmine detection
minas antipersona
MAP
demining technologies
desminado
deteccion de minas
deteccion de explosivos
explosive detection
- Rights
- openAccess
- License
- Atribución-NoComercial 4.0 Internacional
Summary: | Colombia ranks second in the world by number of victims from landmines; and in Colombia, Antioquia is the most affected department. As most landmines in Antioquia do not have metallic parts, metal detectors became useless, however, in most of those mines, the explosive mixture includes ammonium nitrate, and thus a system able to detect this compound could help to locate the mines. On the other side, nuclear quadrupole resonance (NQR) is a spectroscopic technique that allows the detection of some compounds very specifically. Thus, this work had the purpose of developing a system for remote sensing of ammonium nitrate in landmines by NQR. To achieve this goal, a portable NQR device was constructed as well as a probe, capable of sending radiofrequency pulses at the resonance frequency of ammonium nitrate, and capable of picking up the NQR signal after excitation. The manufactured system was tested against some environmental factors, concluding that the most affecting ones are soil conductivity and landmine shape. After achieving a functional system, a classifier based on spectral descriptors was trained, using data acquired with and without ammonium nitrate. Experimental results showed that the proposed classifier (an ensemble of 20 decision trees) had better performance, in terms of the area under the receiver operating characteristic curve, than the classical solution on the literature (which relies only on signal intensity). A final test validated the performance of the system, which detected four of five buried targets in an area of 2 x 1,6 m, having 3 false alarms. |
---|