Cuckoo Search Optimizer based Piecewise Gamma Corrected Auto-clipped Tile-wise Equalization for Satellite Image Enhancement

An efficient optimally weighted framework for piecewise gamma corrected adaptively clipped sectional equalization for overall quality improvement of remotely sensed dark satellite images, is proposed here. In association with it, cuckoo search based biologically inspired optimizer is employed here due to its remarkably attractive exploration and exploitation policies. Tile-wise equalization and associated optimal clipping leads to adaptively equalized interim intensity channel. Piecewise gamma correction using reciprocal gamma values those are derived optimally. Entire allowable intensity span must be exploited so that the count of the void bins can be reduced and hence, cost function is framed by introducing this void-bin count as a penalty term along with desired assurance for entropy as well as contrast enhancement. Rigorous experimentation is executed by employing the performance evaluation and comparison with pre-existing recently proposed and highly appreciated quality enhancement approaches.

I. Introduction

Satellite digital imagery along with its sister applications related to remotely sensed imagery such as scientific anthropological innovations in various sectors of geoscience, astronomy, defense applications, agriculture, mining, weather forecasting, vegetation density analysis, etc., usually rely on the quality and reliability of the captured images so that they can be further relevantly processed [1]. Due to the hardware challenges of the capturing circuitry along its desirable position and unfavorable climatic circumstances, most the times, the acquired images are not up to mark and hence preprocessing is essentially required so that desired quality enhancement can be imparted [2]. Various histogram-based, transform-domain based as well as spatial-domain based strategies [3]–[19] (involving various kinds of intelligence) have been suggested till date for enhancing the visual intensity and gradient based features of the image. Earlier, general histogram equalization (GHE) [20] based enhancement is initially coined for imparting uniform probability distribution. Later onwards, multiple histogram sub-division and local equalization based methods have been also proposed for enhancement of general images. A complete descriptive analysis for these methods is already mentioned in the literature [1]. Various significant contributions like contrast-limited adaptive HE [21] also fascinated the researchers. Intensity level based segmentation and consequently sub-equalized approaches like median-mean based sub-image clipped HE (MMSICHE) [22] has been also suggested. In the same sequence, the averaging HE (AVGHEQ) [23] and later onwards, HE based optimal profile compression (HEOPC) [24] has been proposed for color image enhancement. Focusing on allowable intensity span expansion through count reduction of void intensity bins for color image enhancement has been proposed under the head of histogram equalization with maximum intensity coverage (HEMIC) [25]. For general images, these techniques seem sound; but without imparting gamma correction, it is very tough to impart quality improvement for dark images. Earlier, the adaptive gamma correction with weighting distribution (AGCWD) [26] is introduced, but sometimes leads to the regional over-saturation due to nature of its transformation curve and hence, improved versions [27]–[29] also came to the existence. Other better proposals have been also drafted like the intensity and edge based adaptive unsharp masking filter (IEUMF) [30] by employing the unsharp masking filter which can be further utilized as edge augmentation for imparting overall enhancement. Recently, a piecewise gamma corrected HE (PGCHE) has been also proposed for dark image enhancement. In this paper, a cuckoo search optimizer [31] based piecewise gamma correction is optimally associated with tile-wise equalization followed by adaptive clipping. Remaining content is organized as: Section II presents the core problem formulation. The proposed methodology is available in Section III. Experimentation is discussed in Section IV and finally, conclusion constitutes the Section V.