This page provide an astronomical benchmark data sets.
Matlab routines to read and write in FITS format can be found here.
Astronomical Images
- NGC2997.fits: Galaxy NGC2997 (256x256).
- HaleBopp256.fits: Comet Hale-Bopp (256x256).
- einstein512.jpg : Einstein picture (512x512).
- simu_sky.fits : Simulated HST image (256x256).
- saturn512.fits : Saturn image (512x512).
Gaussian denoising
- bert_orig.fits: Simulated sky image (256x256) using skymaker, and same image + Gaussian noise (sigma=10).
- simu_sky.fits : Simulated HST image (256x256), and same image + Gaussian noise (sigma=50).
Poisson denoising
- sgra_hires.fits: Chandra Galactic Center (2048x2048).
- rosseta.fits: Simulated rosseta image + Poisson noise (512x512), and denoised image f_rosetta_n7_s4_p_A.fits using the method described in (Starck and Pierre, 1998) , and inplemented in the mr_filter program, available at iSAP, with options "mr_filter -m10 -n7 -f3 -k -s4. -v -p -A rosseta.fits f3_rosetta_s3p8_A".
Gaussian Deconvolution
- bpic_data.fits: Infrared image (64x64) of a star (beta pictoris) with a disk, and an observed reference star can be used for the PSF (64x64). Noise follows a Gaussian law (or not too far away).
Poisson Deconvolution
- simu_sky.fits : Simulated HST image (256x256), simulated HST Point Spread Function (PSF) simu_psf.fits , and simulated data simu_data_poisson.fits , i.e. simulated image convolved with the PSF + Poisson noise. Deconvolved image dec_simu_data_poisson_m2_n6_p.fits using the method described in (Starck et al, 1995), and inplemented in the mr_deconv program, available at iSAP, with options "mr_deconv -v -m2 -n6 -p simu_data_poisson.fits simu_psf.fits dec_simu_data_poisson_m2_n6_p.fits".