ABSTRACT
A problem that occurs throughout much of astronomy and other remote sensing applications is how best to interpret noisy data so that the resulting deduced quantities are real and not artefacts of the noise. The process of removing instrumental effects from data can be necessary for any type of measurement, but is perhaps best studied in relation to imaging, when the process is generally known as ‘deconvolution’. Recently, several methods have been developed to aid choosing the ‘best’ version of the deduced quantities from the range of possibilities. The Maximum Entropy Method introduces the external constraint that the intensity cannot be negative and finds the solution that has the least structure in it that is consistent with the data. The alternatives to photography for recording images directly are all electronic in nature. Low-light-level television systems, combined with image intensifiers and perhaps using a slower scan rate than normal have been used in the past.
