Noise-Resistant Remote Sensing System Based on Discrete Tomography Approach and Artificial Intelligence

The investigation of the noise resistance of the system of underground object recognition based on the artificial neural network (ANN) is presented. The surface of the ground with hidden object is irradiated with plane impulse electromagnetic wave. The electromagnetic problem is solved numerically by the method of finite differences in time domain (FDTD). The solution includes all processes of interaction between electromagnetic field, soil, and hidden object. The input data for the ANN are form from time dependences of the received electromagnetic wave. Using discrete tomography approach and ray tracing the additional information about the object is obtained. Utilization of the grid search on problem of modelling parameters of the problem and ANN permits to find the optimal system configurations and time window for successful subsurface object recognition in the presence of noise.