An Inversion Method for Buried Targets Combining BP Network and Outdoor Measurement

Buried object detection is an important application of electromagnetic inverse scattering. However, the inverse scattering problem is usually severely nonlinear and ill-posed. Conventional solutions, such as approximation and optimization methods, are difficult to achieve fast and accurate convergence. As a result, the machine learning method is used to solve those kinds of problems. In the first, simulations and experiments are used to explore the certain laws of radar cross section (RCS) and target depth. Then build a BP neural network to solve inversion problems of object depth based on electromagnetic scattering results. The evaluation shows that the results' root mean square error (RMSE) is only 0.048.