FUZZY METHOD FOR UNDERWATER IMAGE ENHANCEMENT

Abstract

Enhancing images is a fundamental step for most computer vision and other machine learning tasks, ensuring they are ready for subsequent processing. Underwater image enhancement plays a vital role in improving the visual quality of images affected by low light, color distortion, and scattering effects. Underwater haze and images taken in murky underwater environments often exhibit low visibility and poor contrast and complicates the process of retrieving meaningful details from images, often leading to poor image quality. Improving these images by removing haze not only enhances their clarity but also facilitates further analysis. As a result, haze removal is considered both an essential and demanding aspect of underwater image processing. Therefore, enhancing these images is essential for effective analysis during underwater exploration and inspection tasks. Moreover, to be practical for real-time applications, the enhancement methods must be computationally efficient. Histogram Equalization (HE) [1] is the most simple and widely used image enhancing method that enhances the contrast of an image by distributing its pixel intensity values. However it overenhances the image and degrades the quality of the image. Under such premises, my current work focuses on the downside of histogram equalization and closely examines the existing contrast enhancement methodolgies. In this project I explore different existing methods and try to incorporate techniques like Fuzzy logic and Discrete Cosine Transform and see how these algorithms increase the contrast of the Underwater images while also preserving the natural look of the image [19][20].