Deploying Wavelet Transforms in Enhancing Terahertz Active Security Images

Abstract

Clarity of Terahertz images is essential at various security checkpoints to avoid life�s dangers and threats. However, Terahertz images are distorted by noise. During the image gathering, coding, delivery, and processing steps, noise is typically present in the digital image. Without a prior understanding of the noise model, removing noise from images is extremely challenging. Wavelet transforms have gained popularity as a tool for image denoising. In this paper, we advance a solution to this challenge using Global Threshold selection as well as wavelet transform filters. When compared to denoising Gaussian noise at the same percentage induced, biorthogonal is the most effective denoising filter for salt and pepper noise. As the salt and pepper noise increases from 20% to 60%, the hidden security image as our target varnishes or is overpowered by the induced salt and pepper noise. We discover that despite the fact that the bior 4.4 and sym 4.0 wavelet transform filters prove powerful in denoising the image, it is still not clearer and that when an image is tainted by Gaussian noise, wavelet shrinkage denoising is nearly perfect in both bior 4.4 and sym 4.0, whereas when the image is tainted by salt & pepper noise, wavelet shrinkage denoising is nearly perfect in both bior 4.4 and sym 4.0. � 2022, Springer Nature Switzerland AG.