Analisa Perbandingan Metode Arithmetic Mean Filtering dan Metode Konvolusi Pada Citra Bernoise

Abstract

Image capture often experiences problems which can be caused by errors in the camera lens or dirt in the image. This interference is called noise or noise. Noise makes an image unclear and damages the quality of an image. There are many types of noise in images, one of which is salt and pepper noise. This noise will give black and white spots on the image like a sprinkling of salt caused by bit errors when sending data, or damage to the storage area so that filtering is needed on the image to improve the image quality to be better than the original image. The filtering technique used for noisy images is arithmetic mean filter and convolution. Arithmetic mean filtering improves image quality by replacing the pixel value with the average value of its neighboring pixels, while the image convolution technique gives a new value to each pixel by performing several calculation functions from that pixel to the pixels around it. To measure the filtered noise which has decreased, the MSE and PSNR parameters are used. The results obtained through MSE and PSNR which are better used in reducing noise are the arithmetic mean filter. The arithmetic mean filter makes images with salt and pepper noise experience a decrease as seen from the resulting PSNR value which is higher, namely 52% compared to the convolution PSNR value of 47%.