Mamnuah, Mamnuah (2019) Penerapan Metode Dual Tree Complex Wavelet Dan Thresholding Normal Shrink Untuk Reduksi Noise Hujan Pada Video. Other thesis, Institut Teknologi Sepuluh Nopember.
|
Text
06111540000001-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Pada proses pengambilan video terkadang video yang dihasilkan kurang optimal, salah satunya akibat kondisi cuaca yang kurang mendukung seperti terjadinya hujan hujan. Pada penelitian ini hujan di anggap sebagai noise yang mengganggu keaslian dan kualitas video tersebut. Banyak sekali cara untuk mengurangi noise seperti noise hujan, antara lain dengan mendekomposisikan citra dengan transformasi wavelet, kemudian koefisien wavelet di threshold dan direkonstruksi kembali dengan inverse. Adapun pada Tugas Akhir ini dekomposisi frame dari hasil ekstrak video masukan dilakukan dengan metode transformasi Dual Tree Complex Wavelet (DTCWT) dimana pada level pertama menggunakan filter Farras dan filter kedua menggunakan filter Qshift 6 dimana proses perataan nilai matriks dalam domain wavelet pada setiap frame menggunakan thresholding Normal Shrink. Uji coba dilakukan pada video hujan buatan dalam kondisi kamera diam yang terdiri dari objek diam dan objek bergerak. Untuk mengukur kualitas video hasil reduksi, yaitu dengan cara membandingkan nilai PSNR setiap frame hasil reduksi dengan setiap frame pada video sebelum terkena noise hujan. Berdasarkan uji coba dari 6 video didapatkan semakin tinggi nilai standar deviasi (level noise) maka semakin kecil nilai PSNR artinya kualitas video hasil reduksi noise hujan yang diperoleh akan semakin jauh dari video aslinya.
===============================================================================================================================
In the process of recording video, the resulting footage is sometimes not optimal, one reason being unfavorable weather conditions such as rain. In this study, rain is considered a form of noise that disrupts video quality. There are many ways to reduce noise such as rain noise, among others by decomposing the image using wavelet transformation, after which the wavelet coefficients are thresholded and reconstructed again using an inverse transformation. In this final project, the frame decomposition from the input video is performed using the Dual Tree Complex Wavelet (DTCWT) transformation method, where the first level uses the Farras filter and the second level uses the Qshift-6 filter, with the matrix value alignment process in the wavelet domain at each frame using Normal Shrink thresholding. Trials were carried out on artificial rain videos under camera conditions consisting of both stationary and moving objects. From the trials conducted on 6 videos, it was found that the higher the standard deviation value (noise level), the smaller the PSNR value, meaning that the quality of the rain-noise-reduced video obtained becomes further from the original video (the video before rain noise was added).
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | reduksi noise hujan, dual tree complex wavelet, thresholding Normal Shrink |
| Subjects: | T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing |
| Divisions: | Faculty of Mathematics and Science > Mathematics > 44201-(S1) Undergraduate Thesis |
| Depositing User: | - Mamnuah |
| Date Deposited: | 06 Aug 2026 01:12 |
| Last Modified: | 06 Aug 2026 01:12 |
| URI: | http://repository.its.ac.id/id/eprint/69268 |
Actions (login required)
![]() |
View Item |
