Design and Implementation of Real Time Image Retrieval System for De-Noising Space Images in Chandrayaan 3 Lunar Mission Using Autoencoders

In the context of the Chandrayaan 3 Lunar Mission, this research paper introduces a real-time image retrieval and denoising system powered by autoencoders, designed to tackle the challenge of noisy space imagery. Leveraging advanced deep learning techniques, our system employs autoencoders to extract essential features from the noisy lunar images and subsequently retrieves and integrates similar, noise-free reference images from a comprehensive database. By doing so, it achieves real-time denoising, ensuring that the mission’s acquired lunar images are of high quality, thus facilitating more accurate scientific analysis. The paper details the architecture of the autoencoder-based denoising system, its training process using a meticulously curated lunar image dataset, and its seamless integration into the mission’s image processing pipeline. Experimental results underscore the system’s remarkable noise reduction capabilities, thereby playing a pivotal role in enhancing the Chandrayaan 3 mission’s scientific contributions, ultimately advancing our understanding of the lunar environment and bolstering the success of lunar exploration endeavours.