An initial value approach to the inverse Helmholtz problem at fixed frequency

From MaRDI portal





The inverse Helmholtz problem calls for the determination of the function \(f\) in \(\Omega \subseteq \mathbb{R}^n\) from the boundary values \(g(\theta,x) =u(x)\), \(x\in \partial \Omega\) of the solution \(u\) of the Helmholtz equation \[ \Delta u(x)+ k^2 \bigl(1+ f(x)\bigr) u(x)= 0, \quad u(x)= e^{ikx \cdot \theta} +w(x) \tag{1} \] in \(\mathbb{R}^n\), where \(w\) satisfies the Sommerfeld radiation condition. The number \(k>0\) is the frequency of the incident plane wave with direction \(\theta\in S^{n-1}\). We assume \(f=0\) outside \(\Omega\). \(g\) is measured for a single fixed frequency \(k\) and all \(\theta\in S^{n-1}\). The mathematical theory of this inverse problem is fairly well developed. The purpose of the present paper is to give a rigourous mathematical justification of the method already described by \textit{F. Natterer} and \textit{F. Wuebbeling} [(*) Inverse Probl. 11, No. 6, 1225-1232 (1995; Zbl 0839.35146)] and to demonstrate its efficacy by computer simulations. An essential feature in (*) is the use of initial value methods for the Helmholtz equation. We work with initial value problems of second order. The stability analysis is given in \S 2. In \S 3 we describe our method for the inverse problem in a continuous setting. The necessary discretizations have been described in (*). In \S 4 we report the results of the reconstruction of a breast phantom from simulated data. For the entire collection see [Zbl 0871.00041].











This page was built for publication: An initial value approach to the inverse Helmholtz problem at fixed frequency

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4356951)