Thin-layer medium parameter inversion method based on reflection signal reconstruction

Abstract

The invention discloses a thin-layer medium parameter method based on reflection signal reconstruction, which is suitable for a layered medium model containing a thin layer, can solve the time delay estimation problem of the thin layer, is high in precision and convergence rate, can obtain better thin-layer parameter inversion performance, and is an efficient and accurate inversion method. The method is used for inverting underground thin-layer medium parameters under the condition that prior information is unknown, the prior information comprises a detection scene material and actual thickness, and the method comprises the following steps that firstly, a reflection signal model is reconstructed through definition of a generalized reflection coefficient and a convolution model of underground layering, and the model comprises physical parameters of a layered medium; then an optimization problem is constructed, a cost function of the optimization problem is an error square between the model and an actual reflection signal, the optimization problem is solved through a genetic algorithm, and the error square is minimized; and solving an optimization problem through a genetic algorithm, and carrying out optimization in an initial value range.

Description

technical field [0001] The invention relates to the technical field of ground penetrating radar signal processing, in particular to parameter inversion of a layered medium scene containing thin layers. Background technique [0002] GPR is a non-destructive testing tool that uses high-frequency electromagnetic waves to detect features of subsurface structures. The technology has been widely used in shallow subsurface geophysical exploration, such as determining subsurface targets, water content, layer thickness and material density. Layered medium thickness inversion is an important application of ground penetrating radar in practical engineering. It determines the thickness and dielectric parameters of the measurement data based on the reflected signals generated by dielectric discontinuities between different materials. However, limited by the longitudinal resolution of the radar itself, when the layer thickness is very thin, the reflected signals will overlap, making it ...

AI Technical Summary

Problems solved by technology

Inversion methods using optimization algorithms such as gradient descent and Newton iteration can be applied in real-time detection, but the result will fall into a local optimal solution due to too many inversion parameters

[0004] Therefore, it can be seen that in the inversion of ground penetrating radar thin layer medium parameters, due to the aliasing of radar reflection signals in the thin layer, the traditional method cannot solve the problem of time delay estimation of the thin layer well, and some complex algorithms will be due to The amount of calculation is too large to achieve real-time detection

Images