Deconvolution using Meyer wavelets (Q5939750): Difference between revisions

Using Meyer wavelets, the convolution equation \[ \int_{\mathbb{R}} k(t-s) f(s) ds= g(t)\qquad (t\in\mathbb{R}) \] is converted into a well-posed problem in the scaling subspaces provided that \(k\in L^1(\mathbb{R})\) and its Fourier transform \(\widehat k(\omega)\neq 0\). If \(\widehat k\) has a simple zero, the authors look for an approximate solution in both scaling subspaces and wavelet subspaces, which can be used to solve the deconvolution problem numerically. Results related to the convergence rate and error bounds are obtained for Sobolev space functions \(g\). However, for larger resolution level, the resulting discrete system has a larger condition number.