Influence of Noise Reduction on Object Location Classification by Artificial Neural Networks for UWB Subsurface Radiolocation

The practical application of subsurface radars in ultrawideband radiolocation is restricted by a presence of noise in the received signal. The phenomenon creates important limitations on the precision of the data acquired because of low energy of electromagnetic field components generated by subsurface objects of interests in comparison with the incident and reflected from air-ground interface waves. Although the artificial neural networks used to recognize hidden objects have their own abilities to resist the noises and other harmful unpredictable deviations in real measurements of impulse electromagnetic wave the influence of previous denoising is investigated. It is compared the results of object classification of its position by artificial neural network for noisy data of different noise levels and preliminary processed by wavelet transform and caterpillar method of denoising.