Image Enhancement Using Exposure and Standard Deviation-Based Sub-image Histogram Equalization for Night-time Images

Abstract

In this paper, a novel exposure and standard deviation-based sub-image histogram equalization technique is proposed for the enhancement of low-contrast nighttime images. Initially, the histogram of the input image is clipped to avoid the over-enhancement. The clipped histogram is partitioned into three sub-histograms depending on the exposure threshold and standard deviation values. After that, the individual sub-histogram is equalized independently. At last, a new enhanced image is produced after combining each equalized sub-images. The simulation results reveal that our proposed method outperforms over other histogram equalized techniques by providing a good visual quality image. The proposed method minimizes the entropy loss and preserves the brightness of the enhanced image efficiently by reducing the absolute mean brightness error (AMBE). It also maintains the structural similarity with the input image and controls the over-enhancement rate effectively.