Artificial Bee Colony-Optimized Contrast Enhancement for Satellite Image Fusion

Abstract

Image fusion combines two or more images to a single image to extract all the necessary information from the source images. It minimizes the redundant information present in the source images. Fused images find wide applications in medical imaging, computer vision, remote sensing, change detection, and military applications. The success of the fusion technique is limited by the noise present in the source images. In order to overcome this limitations, an artificial bee colony (ABC)-optimized contrast enhancement for satellite image fusion is proposed to fuse two multitemporal satellite images. The ABC-optimized source images are given as input to the fusion stage. A hybrid contrast enhancement technique combining the histogram equalization and gamma correction techniques is used for the contrast enhancement of the source images. The contrast-enhanced images are fused using Discrete Wavelet Transform (DWT), Principle Component Analysis (PCA), and Intensity, Hue, Saturation Transform (IHS) individually. The proposed work further compares these conventional fusion techniques by computing performance measures for image fusion such as Mean Square Error (MSE), Peak Signal-to-Noise Ratio (PSNR), entropy, Structural Similarity Index (SSIM), and Feature Similarity Index (FSIM). The experimental results show that the IHS-based image fusion technique outperforms the PCA- and DWT-based fusion techniques. Also, this method is computationally effective and simple in its implementation.