OPTIMALISASI ANALISIS DENOISING CITRA MENGGUNAKAN ALGORITMA GAUSSIAN FILTER PADA CITRA SATELITE
Daftar Isi:
- Satellite image contains some noise or roar which can pull down a image quality, in order to fix it needs a matrix filtering using gaussian filter that is started Gaussian matrix 3x3, 5x5, 7x7, and 9x9 or higher. This study will find out a matrix on the most optimal Gaussian filter in decreasing noise by seeing MSE and PSNR value from processing image result. A filtering process used 24 data of satellite image that was produced by BMKG of Palembang and was added noise salt & pepper 10 % dan noise specklese as much as 10% for 24 satellite data. This design used static image or not in a real time, then through image processing at cropping stage, noise addition, and filtering, next was an optimalization process used a program to get MSE value, PSNR, and process time on Gaussian matriks 9x9, 90x90, dan 190x19, therefore refraction compared which matrix that was the most optimal to reduce noise on satellite image. . Satellite image that has been done a filtering process used Visual Studio C# 2010 program, the result of the testing was MSE value and PSNR which that value will be lower if Gaussian matrix that was used higher, due to the process of Gaussian filter convolution, edge pixels were ignored and did not convolute. An edge pixels value was same as early image value that has produced a filtering process was only occured in the middle of karnel, hence from the testing of 9x9, 90x90, and 190x190 matrix the most optimal was 9x9 matrix because almost the whole karnel has changed with a new pixels value.