ANALISIS PERBANDINGAN REDUKSI NOISE MENGGUNAKAN METODE MEAN, MEDIAN DAN CONTRA-HARMONIC MEAN FILTERING PADA CITRA GRAYSCALE POLA TENUNAN DAERAH PROVINSI NUSA TENGGARA TIMUR

Abstract

Abstract

The quality of a digital image highly depends on some aspects such as the sharpness level, color sensitivity, and the level of noise. Digital image processing is beneficial to attain better digital image in terms of its quality. This study aims at implementing the noise reduction methods of Mean filtering, Median filtering dan Contra-Harmonic Mean filtering to reduce Gaussian, Speckle dan Salt & Pepper noise on grayscale images of weaving pattern of some regions in East Nusa Tenggara Province. The test is conducted by analyzing the average of Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR) as well as by doing direct visual observation. It shows the contradictive value of MSE and PSNR in which the qualified image gets the lowest value of MSE and the highest value of PSNR. The result of the study confirms that to reduce the gaussian noise, all those three methods gain satisfying results with the PSNR value at around 67 dB, in reducing speckle noise, Mean filtering and Contra-Harmonic Mean filtering methods are quite better compared to Median filtering method with the PSNR value at around 77 dB and to reduce Salt & pepper noise, Median filtering method alone is better than both Mean filtering and Contra-Harmonic Mean filtering method with the PSNR value at 86.2846 dB.

Keywords: mean filtering, median filtering, contra-harmonic mean filtering, gaussian noise, speckle noise, salt & pepper noise