Mine detection by infrared thermography: reduction of size of the experiments

A very promising technique for the detection buried land-mines is infrared thermography. In particular, the presence of a mine can be associated to a specific pattern in the temperature of the soil surface measured by a videothermographic camera. The pattern is induced by the interaction of the daily heating and cooling cycles with the thermal anomaly introduced in the ground by the mine.The specific effectiveness of the thermographic mine detection can be increased by exploiting computerized data-processing tools such as neural-networks, genetic algorithms, or the fusion of dual-band data, to enhance and identify the thermal signature of the mine. The development and ‘training’ of these tools, however, as well as the optimization of the detection equipment and procedure, require a large amount of reference data.Simulations are powerful tools to generate reference data, as they allow fast and inexpensive parametrical analyses. Unfortunately, computer models cannot take care of all the variables influencing a real-world problem and, furthermore, they need a proper calibration. On the other hand, performing parametrical analyses by on-field experiments can be long and challenging, since in an outdoor test environment and along the 24-hour long day period it is arduous to control selectively the phenomena that influence the thermal problem, especially the weather conditions.In view of this, the main objective of the present work is to develop a method by which the cycles of heating and cooling of the soil can be replicated in the laboratory with a reduced scale, either in time or in space, but obtaining measures that can be directly correlated to realistic on-field tests. The result is sought through the dimensional analysis of the equations governing the thermal problem, and it is verified by computer simulations and experiments