A Novel Approach for Identifying and Eliminating the Degradations in Real Time Images
Year: 2025
Type: article
Abstract: Satellite imagery plays a crucial role in various applications, but the quality of the images can be degraded by noise. Accurate noise detection is essential for effective image enhancement and analysis. This project proposes a hybrid approach for noise detection and enhancement in satellite images by integrating deep learning and conventional techniques. We use a convolutional neural network (CNN) model along with Daubechies wavelet coefficient and statistically extracted coefficient paired with feature components extracted after applying median and high-pass filter for noise detection, to classify the type of noise in satellite images. The CNN model is trained on a dataset of images with various noises, learning to discriminate between different noise patterns. After classifying the type of noise, we apply a enhancement approach utilizes deep learning techniques by leveraging U-Net based CNN model to reduce noise levels in an image. Our approach achieves a validation accuracy of 91.29% for noise classification and a PSNR of 37.78 dB for image enhancement, outperforming several state-of-the-art methods.
Source: IEEE Access
Authors: S. Rajkumar, L. M. Jenila Livingston, Vansh Juneja, Shashwat Sharv, Pratik Dattatray Mahajan
Institution: Vellore Institute of Technology University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,075
OpenAlex ID: W4411358759