Blind deconvolution for astronomical spectrum extraction from two-dimensional multifiber spectrum images.

Although advances in astronomical observation techniques and fiber optic technology have enabled two-dimensional (2D) multifiber spectrum images, astronomers often need one-dimensional (1D) spectroscopy to study physical and chemical properties of astronomical objects. Because of faint optical flux and light pollution, determining point spread functions (PSF) for large-scale multifiber spectroscopic telescopes is difficult. We propose a new optimal extraction method that uses blind deconvolution to extract 1D astronomical spectroscopy from 2D multifiber spectrum images without knowing the exact PSF. A comparison of the performance of our blind deconvolution extraction method and those of other extraction methods showed consistent results, indicating that the blind deconvolution extraction methodology is useful in analyzing 2D multifiber spectrum images and reducing fiber-to-fiber crosstalk without degrading the spectrum's resolution.