Pengenalan Tekstur Pahatan Pada Citra Prasasti Menggunakan Backpropagation

Chairy, Amalia (2017) Pengenalan Tekstur Pahatan Pada Citra Prasasti Menggunakan Backpropagation. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
2215205004-Master_Thesis.pdf - Accepted Version

Download (8MB) | Preview

Abstract

Prasasti adalah dokumen penting yang diwariskan oleh sejarah kerajaan. Prasasti terbuat dari bahan keras seperti batu dan tembaga. Oleh karena itu perlu dilakukan digitalisasi dokumen, untuk menjaga keaslian dokumen. Namun dokumen warisan sejarah memiliki gangguan pada plat prasasti yang disebut noise. Sehingga perlu dilakukan pengurangan noise pada gambar prasasti. Salah satunya dengan memisahkan latar belakang dan objek tulisan yang diukir diatas prasasti agar memudahkan untuk dibaca. Ekstraksi ciri merupakan satu hal penting untuk dilakukan dalam pengolahan gambar, Karena dari hasil ekstraksi ciri bisa diperoleh informasi penting mengenai karakteristik gambar tersebut. Salah satu ciri yang bisa dianalisis adalah ciri tekstur. Pada penelitian ini menggunakan prasasti asal Indonesia yang bernama prasasti Adan-adan. Prasasti ini tersimpan di museum Mpu Tantular, Jawa Timur, Indonesia. Prasasti yang memiliki patina berwarna coklat. Kemudian diekstraksi dengan menggunakan metode ekstraksi fitur tekstur orde pertama dan kedua. Dari hasil ekstraksi, fitur tekstur yang digunakan adalah gray level Co-occurrence Matrix(GLCM). Dimana terdapat tiga fitur yang bisa digunakan, yaitu IDM, Korelasi, dan Entropi. Setelah dilakukan ekstraksi fitur tekstur, dilakukan pelatihan dan pengujian untuk mengenali pahatan pada citra prasasti menggunakan metode Backpropagation. Sehingga hasil yang diperoleh ada 89.85% citra yang dikenali dengan menggunakan metode backpropagation. ======================================================================================================= Inscription is an important document inherited by history of kingdom. Inscription made on hard stuff such as stone and copper. There fore it is necessary digitizing documents, to keep the authenticity of the document. But the document of thehistorical heritage have disruption on inscription plate which be called noise. So that, it is necessary to reduce the noise in the image of the inscription, to ease the documentation of historical digital. Then, separation between the background and the writing object carved on inscription is conducted so easy to read. Feature extraction is one important thing to do in image processing, Because of the feature extraction results, important information about the characteristics of the image can be obtained. One of the features that can be analyzed is a feature of texture. This paper uses an inscription image. Inscription is an important document inherited by the history of kingdom. This paper uses an inscription from Indonesia called Adan-adan. This inscription is stored in the museum of Mpu Tantular, East Java, Indonesia. The inscription has a brown patina, then extracted by using texture feature extraction method of first and second order. Patina is a fine layer of oxide on the surface of the metal / copper. From the extraction results, the texture feature used is gray level Co-occurrence Matrix (GLCM). There are three features that can be used, namely IDM, Correlation, and Entropy. After the extraction of texture features, training and testing recognize sculpture on the image inscriptions using backpropagation method where conducted. As a results obtained, 89.85% of images was recognizable by using Backpropagation method.

Item Type: Thesis (Masters)
Additional Information: RTE 621.367 Cha p-1 3100018074119
Uncontrolled Keywords: Ekstraksi Ciri; Prasasti; Dokumen sejarah; Gray level Co-occurrence Matrix(GLCM); Backpropagation; Fea ture Extraction; Inscription; historical documents; Gray -level Co-occurrence Matrix (GLCM)
Subjects: Q Science > Q Science (General) > Q325.78 Back propagation
T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
Divisions: Faculty of Electrical Technology > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User: Amalia Chairy
Date Deposited: 12 Feb 2018 03:20
Last Modified: 27 Apr 2020 23:59
URI: http://repository.its.ac.id/id/eprint/49235

Actions (login required)

View Item View Item