ABSTRAKSI: Saat ini citra digital telah dapat dimanfaatkan dalam berbagai bidang kehidupan. Hal ini menyebabkan akses dan perpindahan citra digital dari suatu media ke media lain semakin sering dilakukan. Akan tetapi sering kali dalam proses perpindahan tersebut citra digital mengalami distrosi yang mengakibatkan adanya noise pada citra digital yang diterima, sehingga menyebabkan kualitas citra yang diterima menjadi berkurang atau tidak sesuai dengan citra aslinya.
Dalam tugas akhir ini telah diimplementasikan, dianalisis dan dibandingkan pemakaian dua buah metode yaitu Switching Median Filter With Boundary Discriminative Noise Detection dan Adaptive Median Filter dalam melakukan proses pengurangan noise pada suatu citra digital, sehingga kualitas citra dapat ditingkatkan. Noise yang dipakai adalah jenis impulsive noise dengan probabilitas tertentu yang dibangkitkan melalui suatu noise generator.
Parameter performansi yang diujikan pada citra digital adalah PSNR (Peak Signal-to-Noise Ratio) dan MOS (Mean Opinion Score) pada citra hasil filtering. Performansi kedua metode tersebut akan dibandingkan dengan performansi dari Standard Median filtering yang umum digunakan.
Dari hasil analisis didapatkan bahwa secara umum metode Switching Median Filter With Boundary Discriminative Noise Detection dan Adaptive Median Filter sangat layak digunakan untuk mengurangi impulsive noise dengan nilai probabilitas noise antara 0 sampai dengan 0,4 dengan kenaikan 0,1. Metode Switching Median Filter With Boundary Discriminative Noise Detection sangat baik untuk mengurangi impulsive noise pada citra yang memiliki karakteristik low contrast dan normal brightness sedangkan metode Adaptive Median Filter sesuai untuk megurangi impulsive noise pada citra dengan karakteristik Normal dan High contrast disertai normal brightness.
Kata Kunci : Switching Median Filter With Boundary Discriminative Noise Detection, Adaptive Median Filter, Standard Median Filtering, Noise, PSNR, MOSABSTRACT: Nowdays, digital image can be used in every field of live. It causes access and move from one media digital image to another media more often to do. But, the move often realizes a distortion that makes there is a noise at the digital image that someone accept, so the qualities that accept by someone will be less or not as good as the original.
In this final project, had been implemented, analyted, and compared two methode, Switching Median Filter With Boundary Discriminative Noise Detection and Adaptive Median Filter in doing filtering process to a digital image that may subtract the noise, so the qualities can be advance. Noise that’s used is impulsive noise with a fixed probability, where it’s generated by a noise generator.
Performance parameter that’s tested is PSNR (Peak Signal-to-Noise Ratio) and MOS (Mean Opinion Score) in image coming from the result of filtering. Both of their’s performance will be campare with a performance from Standard Median Filter that usually used.
From the analysis result, asserts that Switching Median Filter With Boundary Discriminative Noise Detection and Adaptive Median Filter method is very proper to be used for decreasing impulsive noise by noise probability value between 0 until 0,4 with increment 0,1. Switching Median Filter With Boundary Discriminative Noise Detection method is very powerful to reduce impulsive noise on digital image that have low contrast and normal brightness characteristic whereas Adaptive Median Filter is very appropriate to reduce impulsive noise on digital image that have Normal and High contrast with normal brightness characteristic.
Keyword: Switching Median Filter With Boundary Discriminative Noise Detection, Adaptive Median Filter, Standard Median Filtering, Noise, PSNR, MOS