Detection of groundwater by ground penetrating radar

The application of ground probing radar (GPR) to detection of groundwater from relatively deep aquifers in a desert environment is addressed and processing techniques to improve the detectability of a weak signal in noise and interference are reported. The study is based on simulated images from structures that have the potential of storing groundwater one of which is the buried valley structure. To increase the signal to noise ratio to achieve a reasonable probability of detection and false alarm, various processing schemes are possible, typically employing analogue, binary (double threshold) and digital processing. Different system architectures are compared to improve detectability. Automatic detection and classification by artificial neural networks is tried to classify geologic subterranean features for aiding and speeding the process and to overcome lack of experts on the field.