Peningkatan Kualitas Citra Underwater Menerapkan Metode Difusi Nonlinier Anisotropik

Abstract

Underwater images are often influenced by unwanted noise. Noise in underwater images can make it difficult to take pictures underwater, so noise needs to be reduced. Reducing noise in underwater images often has a blurry effect on all parts of the image. This underwater image is very important, especially taking pictures under water because the underwater image shows the boundaries of the part where the picture is taken underwater. To get the best noise reduction in this research, several trials were carried out. The stages of this method are divided into several stages: Carrying out a Gaussian filter, calculating the diffusivity scale, and comparing the results of the Underwater Image with PSNR. The testing process by identifying Underwater image test data using the Anisotropic Nonlinear Diffusion Method produces PNSR which significantly reduces noise and produces very good images. The results of the PNSR value obtained a decibel (dB) value of 34.3173. These results identified the corrected underwater image samples as having good quality because they were above the regulatory limit of 30 decibels (dB). Where the assessment if below 30 dB indicates that the quality is relatively low because the insertion distortion is very visible