Maliki, Ahmmad (2019) Denoising Citra Radiografi Menggunakan Modifikasi Difusi Anisotropik Untuk Mendeteksi Cacat Las. Sarjana thesis, Universitas Brawijaya.
Abstract
Pengelasan merupakan proses menyambungkan logam padat dengan cara mencairkan logam tersebut melalui pemanasan. Cacat las kecil pada citra radiografi yang mengandung noise sulit dideteksi pada Non-Destructive Test(NDT) sehingga perlu dilakukan denoising. Pada skripsi ini dilakukan denoising citra radiografi untuk mendeteksi cacat las menggunakan modifikasi difusi anisotropik. Persamaan difusi anisotropik merupakan salah satu persamaan diferensial parsial parabolik yang digunakan dalam pengolahan citra digital khususnya pada image filtering. Pada skripsi ini dilakukan modifikasi difusi anisotropik yaitu pada koefisien difusi. Kriteria pemberhentian iterasi juga diperkenalkan untuk mengakhiri proses difusi dengan memanfaatkan entropi citra. Modifikasi koefisien difusi memanfaatkan peluang tingkat keabuan lokal yang dihitung pada daerah ketetanggaan 3
English Abstract
Welding is the process of connecting two solid metals by melting the two metals through heating. Small weld defects in radiography images containing noise are difficult to be detected in Non- Destructive Test (NDT) so denoising is necessary. This final project proposes denoising radiography images to detect weld defects using a modified anisotropic diffusion. Anisotropic diffusion equation is one of the parabolic partial differential equations used in digital image processing, especially in image filtering. This final project proposes modified anisotropic diffusion in the diffusion coefficient. The stopping criterion is also introduced to terminate the diffusion process by utilizing image entropy. Modification of the diffusion coefficient utilizes local gray level probability which is computed in the 3 × 3 neighborhood region of a pixel. To obtain a numerical scheme, discretization of anisotropic diffusion equations is carried out using finite difference method Forward in Time Centered in Space (FTCS). Then stability analysis is carried out on the obtained scheme using the von Neumman method. This model is compared to existing models such as the classic P-M Model and CTAD Model. Image noise is artificial noise with three types of noise, namely Gaussian noise, salt and pepper noise, and speckle noise. Evaluation tools used are MSE, Entropy, PSNR, and computation time. The comparison results show that the modification of the Model P-M has better performance compared to the other two models
Other obstract
-
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/MIPA/2019/430/052001564 |
| Uncontrolled Keywords: | cacat las, denoising citra, persamaan difusi anisotropik, metode beda hingga Forward in Time Centered in Space (FTCS), weld defects, image denoising, anisotropic diffusion equation, finite difference method Forward in Time Centered in Space (FTCS). |
| Subjects: | 500 Natural sciences and mathematics > 519 Probabilities and applied mathematics > 519.5 Statistical mathematics > 519.55 Time-series analysis |
| Divisions: | Fakultas Matematika dan Ilmu Pengetahuan Alam > Matematika |
| Depositing User: | Budi Wahyono Wahyono |
| Date Deposited: | 10 Aug 2020 07:31 |
| Last Modified: | 10 Aug 2020 07:31 |
| URI: | http://repository.ub.ac.id/id/eprint/180002 |
Actions (login required)
 |
View Item |