Survey on Enhancement Techniques of Underwater Images

Abstract

In underwater view, the light which is dependent on wavelength incorporates and scatters the disgrace visibility of images, because of less difference and unclear color casts. To deal with this crisis, a neural network-based image enhancement model, which is skilled expertly by underwater image database. With underwater imaging models and optical properties of underwater view, first combine ten different aquatic image databases. Some of the enhancement algorithms are used to improve various features of underwater images. While putting forward a new algorithm, compare the projected performance with a few substitutes which is there in present situation. This paper gives the valuable methodology definition for underwater image enhancement technique’s effective performance estimation. The images are enhanced by three different approaches:
• Objective metrics,
• An expert panel and
• Estimation of 3D reconstructions results.