Perbandingan Metode Run-Length Dan Co-Occurrence Untuk Ekstraksi Ciri Dengan Metode K-Mean Untuk Mengenali Objek Berdasarkan Tekstur

Hirdan, Rivan Rahmadditya (2015) Perbandingan Metode Run-Length Dan Co-Occurrence Untuk Ekstraksi Ciri Dengan Metode K-Mean Untuk Mengenali Objek Berdasarkan Tekstur. Other thesis, Universitas Komputer Indonesia.

[thumbnail of 1. UNIKOM_Rivan Rahmadditya Hirdan_COVER.pdf] Text
1. UNIKOM_Rivan Rahmadditya Hirdan_COVER.pdf

Download (55kB)
[thumbnail of 4. UNIKOM_Rivan Rahmadditya Hirdan_KATA PENGANTAR.pdf] Text
4. UNIKOM_Rivan Rahmadditya Hirdan_KATA PENGANTAR.pdf

Download (194kB)
[thumbnail of 5. UNIKOM_Rivan Rahmadditya Hirdan_DAFTAR ISI.pdf] Text
5. UNIKOM_Rivan Rahmadditya Hirdan_DAFTAR ISI.pdf

Download (11kB)
[thumbnail of 10. UNIKOM_Rivan Rahmadditya Hirdan_DAFTAR PUSTAKA.pdf] Text
10. UNIKOM_Rivan Rahmadditya Hirdan_DAFTAR PUSTAKA.pdf

Download (86kB)
[thumbnail of 11. UNIKOM_Rivan Rahmadditya Hirdan_BAB 1.pdf] Text
11. UNIKOM_Rivan Rahmadditya Hirdan_BAB 1.pdf

Download (215kB)
[thumbnail of 12. UNIKOM_Rivan Rahmadditya Hirdan_BAB 2.pdf] Text
12. UNIKOM_Rivan Rahmadditya Hirdan_BAB 2.pdf

Download (518kB)
[thumbnail of 14. UNIKOM_Rivan Rahmadditya Hirdan_BAB 4.pdf] Text
14. UNIKOM_Rivan Rahmadditya Hirdan_BAB 4.pdf

Download (425kB)
[thumbnail of 15. UNIKOM_Rivan Rahmadditya Hirdan_BAB 5.pdf] Text
15. UNIKOM_Rivan Rahmadditya Hirdan_BAB 5.pdf

Download (180kB)
[thumbnail of 18. UNIKOM_Rivan Rahmadditya Hirdan_LEMBAR PENGESAHAN.pdf] Text
18. UNIKOM_Rivan Rahmadditya Hirdan_LEMBAR PENGESAHAN.pdf

Download (235kB)
[thumbnail of 19. UNIKOM_Rivan Rahmadditya Hirdan_SURAT KETERANGAN TIDAK PLAGIAT.pdf] Text
19. UNIKOM_Rivan Rahmadditya Hirdan_SURAT KETERANGAN TIDAK PLAGIAT.pdf

Download (39kB)
[thumbnail of 22. UNIKOM_Rivan Rahmadditya Hirdan_SURAT HAK EKSLUSIF.pdf] Text
22. UNIKOM_Rivan Rahmadditya Hirdan_SURAT HAK EKSLUSIF.pdf

Download (206kB)

Abstract

Pengenalan tekstur cukup memegang peranan penting dalam pengolahan citra digital karena dikembangkan dengan tujuan agar komputer dapat memahami serta mengenali tekstur sama seperti yang dilakukan oleh mata manusia. Karena komputer tidak memiliki indra penglihatan, maka komputer hanya dapat mengenali citra digital berdasarkan ciri atau karakteristik teksturnya. Beberapa metode untuk memperoleh ciri-ciri tekstur adalah metode co-occurrence dan metode run-length. Ciri-ciri tekstur yang didapat dari metode matriks co-occurrence diantaranya adalah entropi, kontras, homogenitas dan energi. Sedangkan pada metode run-length adalah short run emphasis (SRE), long run emphasis (LRE), graylevel uniformity (GLU), run-length uniformity (RLU) dan run percentage (RPC). Dari hasil ciri-ciri tersebut kemudian digunakan untuk klasifikasi dengan menggunakan metode k-mean yang menentukan hasil klasifikasi berdasarkan nilai jarak terkecil. Objek yang diuji adalah kayu, pasir, logam dan beras.Metode ekstraksi ciri co-occurrence dan metode run-length memiliki pengaruh dalam hasil klasifikasi yang dilakukan dengan metode k-mean, hal ini terlihat dari tingkat akurasi dan waktu yang diperlukan. Metode co-occurrence memiliki rata � rata akurasi sebesar 87% dengan rata � rata waktu eksekusi 0,7 s sedangkan metode run-length memiliki rata � rata akurasi sebesar sebesar 87,5% dengan rata � rata waktu eksekusi 0,7 s. Berdasarkan pengujian didapatkan metode co-occurrence dan metode run-length memiliki keefektifan tinggi dalam mengenali citra kayu, logam, pasir maupun beras

Texture recognition has an important role in digital image processing because it was developed with the aim for computer to understand and recognize the same texture as same as the human eye. Because computers don�t have the sense of sight, then the computer can only recognize the digital image texture based on traits or characteristics. Several methods for obtaining the characteristics of the texture is co-occurrence method and run-length method. Texture characteristics obtained from the co-occurrence matrix method are entropy, contrast, homogeneity and energy. While on the run-length method are short run emphasis (SRE), long run emphasis (LRE), graylevel uniformity (GLU), run-length uniformity (RLU) and run percentage (RPC). From the result of these characteristics then use for classification using k-mean method which is determines the classification result based on the value of the smallest distance. The object that being tested are wood, sand, metal and rice. Feature extraction of co-occurrence method and run-length method has some influence on the results of the classification performed by the k-mean method, it can be seen from the level of accuracy and the time required. Method of co-occurrence has average accuracy is 87% with the average execution time is 0.7s while the run-length method has average accuracy is 87,5% with th average execution time is 0.7s. based on the obtained test co-occurrence method and run-length method has a higher effectiveness in recognizing the image of wood, sand, metal and rice.

Item Type: Thesis (Other)
Additional Information: Migrasi dari GDL. Identifier: jbptunikompp-gdl-rivanrahma-33424. Folder: S1-Final_Project / Fakultas_Teknik_dan_Ilmu_Komputer / Teknik_Informatika / 2015
Uncontrolled Keywords: tekstur, ekstraksi ciri, co-occurrence, run-length, klasifikasi, k-mean
Divisions: S1 Skripsi > Teknik Informatika
Depositing User: Robot Migration
Date Deposited: 05 Nov 2026 09:10
Last Modified: 05 Nov 2026 09:10
URI: https://elibrary.unikom.ac.id/id/eprint/38203

Actions (login required)

View Item
View Item