H. Parsiani, M. Torres, and P. Rodrguez (Puerto Rico)
: Radar, Material Characteristics, NeuralNetwork, Fourier Transform
The ground penetrating radar (GPR) incident wave to ground is not totally reflected, part of it travels to the next layer, and its intensity and form depends on the boundary reflectivity and the type of materials in the experiment. It is observed that the reflected wave is attenuated, convolved, and compressed differently, as it travels through distinct materials. A material characteristic in Fourier domain (MCFD) is defined and calculated at every reflection. It is possible to determine the characteristics of the media by using the wavelet of the electromagnetic signal before and after it is reflected from the media. An algorithm is developed which calculated MCFD trains a 2-layer Back propagation Neural Network (NN). This NN is further linked to second NN that determines the material moisture. Material type can be determined irregardless of the layer at which the object is buried (limited to GPR reflection intensity level), object size, or if an extended subsurface layer is present. It was also found that by using the MCFD, vegetation could be classified according to its health using GPR. This is a great advantage in terms of higher resolution achievable by active as opposed to passive satellites in the measurements of vegetation health index.
Important Links:
Go Back
