Relating soil properties to performance of metal detectors and ground penetrating radars

This chapter first reviews the basic soil properties that affect the detection performance of metal detectors and ground-penetrating radars, and then a number of soil classes are introduced. A crucial point of the soil class definition is that the class can be assessed as a function of measurable effects on the performance. We show how a database containing soil parameters and corresponding soil classes can be used to train a classifier. With such a classifier, making soil measurements is enough to assess the difficulty of a soil. It is not necessary to first select detectors and make specific performance tests with them. With the classifier, information on soil difficulty can thus be obtained with fewer measurements and the result is more general as it is not related to a single detector. The soil class can be used to plan de-mining operations rationally because it allows the planner, with few measurements, to define soil difficulty and choose the most suitable detector. No database is currently available. We provide guidelines on what the database should contain. Nevertheless, to illustrate the soil classification concept, a simple metal detector model is developed and used to simulate measurements, which are then used to train a classifier.