ID-KOS Method: Image Defogging via Kookaburra Optimized Stationary Wavelet Decomposition Method
B. Ben Sujitha, Mohamed M. Hassan, A. Ahilan, Rashmi Sanjay, K. Selvi, Bajarang Prasad Mishra · IETE Journal of Research · 2025
The defogging of images in surveillance systems has gained much attention in recent years to restore hazy or foggy images. It is difficult to integrate image depth of detail and color into defogging algorithms. In this research, a novel ID-KOS approach which is a combination of the Kookaburra Optimization algorithm and stationary wavelet decomposition for efficient defogging. Initially, the input fog images are input to the Stationary Wavelet Transform (SWT) is separated into high and low frequency sub-images. Then, the low-frequency images are denoised using Multiscale Retinex with color restoration (MSRCR) to remove the fog, and high-frequency images are processed using Curvelet Transform to remove the noise. Kookaburra Optimization Algorithm (KOA) actively enhances color properties and MSRCR filters low-frequency illumination. A curvelet denoising filter and the Unsharp masking (USM) algorithm are used to eliminate high-frequency noise in images. Finally, an inverted SWT algorithm is used to reconstruct an image that is noise-free from sub-enhanced images. According to the experimental results, the proposed approach performs higher than existing techniques when utilizing the established dataset for both quantitative and qualitative assessment criteria. Additionally, the proposed approach effectively eliminates fog without compromising the authenticity of fog images.