For an introduction to this command please see Deconvolution Tutorial.
This command performs either Maximum Entropy or Lucy-Richardson deconvolution of an image. Deconvolution is an advanced image restoration technique which can remove the effects of blurring in an image. This technique, first pioneered for radio astronomy, became very important for visible-light astronomers when the problems with the Hubble Space Telescope were discovered. Even after the installation of corrective optics on the Hubble's instruments, deconvolution techniques have remained important and widely used tools for processing and improving images both from HST and from ground-based observatories.
Outside of astronomical applications, the same image techniques can be applied to just about any image, from microscope pictures to security camera video frames. The only essential requirement is that the image be blurry! Undersampled images, such as those taken with very short focal lengths, will already be as sharp as is possible. If the pixels are too big to resolve the blur, then deconvolution cannot improve the resolution.
Only limited image processing should be performed before deconvolution. CCD images should be calibrated first to minimize consistent effects introduced by the camera (gain, thermal noise) and optical system (flat fielding). If the camera does not produce square pixels, it is recommended that you interpolate them to a square aspect ratio before proceeding. Hot and dead pixels should be fixed. If you do not wish to deconvolve each channel separately, color combining of RGB or CMY images may be done beforehand; for LRGB and LCMY images, you will achieve better results by deconvolving the luminance image before color combining. It is strongly recommended that all other processing functions be deferred until after deconvolution.
The success of deconvolution depends critically on the Point-Spread Function and Noise models, which must be initialized before you can perform the processing itself. If the selected noise model is Poisson, or when using the Lucy-Richardson method, the correct Photoelectrons per ADU setting must be entered; this can be done within the Deconvolve dialog, or separately using the Photons Wizard. Images acquired by MaxIm DL already include this parameter in the FITS header for many camera plug-ins. Settings of all these parameters persist within a particular MaxIm DL session, but you must be careful to avoid processing a mix of images from different cameras or telescopes without resetting them for each different configuration.
Since experimentation is sometimes required to determine the settings which will produce the best results, a mechanism is provided to allow quick evaluation by operating on a subset of the image.
Availability of this feature depends on Product Level.
The Deconvolve command is implemented as a tabbed dialog box with tabs for each of the three steps in the process:
Specifying the Noise Model
Specifying the PSF Model