UXO Signal Multi Sensor Detection and Estimation

In this chapter, the original advanced algorithms for stepped-frequency GPR imaging are considered. In stepped-frequency GPR, the range profile formation is carried out by reconstruction of a wideband chirp by combining a set of stepped-frequency chirp signals in the time domain. Using the Modelsim simulator, it is shown that the processor VIRTEX II Pro is suitable for implementation of this algorithm. A simple convolution algorithm for simulation of stepped-frequency GPR images from multi-layered subsurface media is described. Different approaches and algorithms for the basic GPR signal and image processing are also considered in this chapter. These algorithms are used for improving the image quality of underground objects, e.g. pipes. It is shown that applying different filters (CFAR, Hough, Kalman, Particle) to GPR image processing is a good decision in the sense of estimation accuracy, probability of target detection and false alarm.