Image enhancement based on information entropy

Histogram Equalization (BE) is image enhancement arithmetic that realizes histogram transform through point operation of gray level. It can expand dynamic range of gray level of image and improve vision effect of image. However, Information entropy is proposed to analyze physical significance of HE and disadvantage of standard HE is pointed out. Histogram distribution of image reflects information capacity to a certain extent. HE that is optimal processing making entropy tend to maximum drives lots of gray levels with less probability distribution to be incorporated without considering discrete feature of gray level of image. So part of detailed information with high frequency is lost, then entropy decrease is produced. Therefore two solutions that prevent gray level incorporation in maximum are put forward. Those are HE extending gray level range and improved histogram modified arithmetic. In the former the original grey level numbers are extended to preserve enough resource, hence the incorporation for the grey level with less probability is avoided; In the latter human eye feature is considered to modify the standard HE, Therefore, the intervals between the grey levels with more probability are driven to moderate to save part resource for the grey levels with less probability. Consequently, the incorporation for the grey level with less probability is also prevented.. Experiment results show that grey levels incorporated are much less than standard HE in both two methods, so the vision effect is improved enormously.