Low-pass filtering applied to photon-limited images: improvement by histogram specification
Pedro M. Prieto, Manuel P. Cagigal · Pure and Applied Optics Journal of the European Optical Society Part A · 1995
Low-pass filtering is an already known technique used to smooth the noise in digital images. It has the advantage of simplicity, but, however, it produces poor results when it is applied to photon-limited images. In this paper we show that the histogram of all the post-filtering images is heavily dependent on the statistics of the detected photopulses. Hence, when this dependence is suppressed, the image quality is improved. The solution we propose is a histogram specification based on a fitting procedure to provide the recovered image of a histogram similar to that of the original image. The technique has been applied to very low light level simulated images (about 500 photopulses in a 359*175 pixel frame) and fairly good results have been obtained.