Abstract:This paper is concerned with the problem of inverse scattering of time-harmonic acoustic plane waves by a two-layered medium with a locally rough interface in 2D. A direct imaging method is proposed to reconstruct the locally rough interface from the phaseless total-field data measured on the upper half of the circle with a large radius at a fixed frequency or from the phased far-field data measured on the upper half of the unit circle at a fixed frequency. The presence of the locally rough interface poses challenges in the theoretical analysis of the imaging methods. To address these challenges, a technically involved asymptotic analysis is provided for the relevant oscillatory integrals involved in the imaging methods, based mainly on the techniques and results in our recent work [L. Li, J. Yang, B. Zhang and H. Zhang, <a class="link-https" data-arxiv-id="2208.00456" href="https://arxiv.org/abs/2208.00456">arXiv:2208.00456</a>] on the uniform far-field asymptotics of the scattered field for acoustic scattering in a two-layered medium. Finally, extensive numerical experiments are conducted to demonstrate the feasibility and robustness of our imaging algorithms.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the inverse scattering problem of reconstructing a two - layer medium with a locally rough interface in two - dimensional cases, through total - field data with missing phase or far - field data with missing phase. Specifically, the paper proposes a direct imaging method to reconstruct the locally rough interface from the total - field data with missing phase on a large semicircle at a fixed frequency or the far - field data with missing phase on the upper half of the unit circle.
### Specific description of the problem
1. **Background and challenges**:
- The research object is a two - layer medium system composed of two unbounded media with different physical properties, and its interface is a locally rough surface.
- The existence of the locally rough interface makes the theoretical analysis complex, especially when dealing with scattered waves and reflected waves.
- In many practical applications, it is much more difficult to obtain the phase information of the wave field than to obtain the intensity (or modulus) information. Therefore, it is of great significance to study the inverse scattering problem of data with missing phase.
2. **Existing work**:
- Many numerical algorithms have been proposed to recover impenetrable or penetrable locally rough surfaces from scattered - field data or far - field data.
- Recent work has mainly focused on using techniques such as multi - frequency iterative methods, linear sampling methods, and reverse - time migration methods to deal with impenetrable locally rough surfaces.
- However, relatively few studies have been carried out on recovering penetrable locally rough surfaces from data with missing phase.
3. **Contributions of this paper**:
- Two non - iterative numerical methods are proposed for reconstructing the locally rough interface from total - field data with missing phase and far - field data with missing phase respectively.
- Through complex asymptotic analysis of relevant oscillatory integrals, the theoretical analysis problems brought by the two - layer medium background are solved.
- Detailed numerical experiments are provided to verify the effectiveness and robustness of the proposed algorithms.
### Mathematical model and method
- **Scattering problem**: Consider the scattering problem of a time - harmonic acoustic plane wave in a two - layer medium, where the incident wave is \(u_i(x, d)=e^{ik_+x\cdot d}\), and the total field is \(u_{\text{tot}}(x, d)=u_0(x, d)+u_s(x, d)\), where \(u_0(x, d)\) is the reference wave and \(u_s(x, d)\) is the scattered wave.
- **Boundary conditions**: On the locally rough interface, the transmission conditions \([u_{\text{tot}}]=0\) and \([\partial u_{\text{tot}}/\partial\nu]=0\) are satisfied.
- **Far - field asymptotic behavior**: The scattered wave \(u_s(x, d)\) has the asymptotic behavior \(u_s(x, d)=\frac{e^{ik_+|x|}}{\sqrt{|x|}}u^\infty(\hat{x}, d)+u_s^{\text{Res}}(x, d)\) in the far field, where \(u^\infty(\hat{x}, d)\) is the far - field pattern and \(u_s^{\text{Res}}(x, d)\) is the residual term.
### Imaging function
- **Imaging function for total - field data with missing phase**:
\[
IP(z, R)=\int_{\partial B_R^+}\left|\int_{S_1^-}\left[|u_{\text{tot}}(x, d)|^2-\left(1 + |R_0(\theta_d)|^2+R_0(\theta_d)e^{2ik_+x_2d_2}\right)\right]e^{ik_+ (x - z)\cdot d}-e^{ik_+ (x' - z')\cdot d}\,ds(d)\right|^2\,ds(x)
\]
- **Imaging function for far - field data with missing phase**:
\[
IF(z)=\int_{S_1^+}\left|\int_{S_1^-}u^\infty(\hat{x}, d)\right|