Klasifikasi Citra Satelit Menggunakan Kombinasi Fitur Warna Dan Fitur Tekstur

Penulis

  • Sutrisno .
  • Ahmad Afif Supianto

DOI:

https://doi.org/10.25126/jtiik.201522141

Abstrak

Abstrak

Penelitian tentang klasifikasi citra satelit untuk mengklasifikasikan citra dalam kelompok tertentu sedang mengalami perkembangan. Terdapat masalah yang disebabkan oleh kesalahan yang dilakukan saat ekstraksi fitur. Pada penelitian ini, peneliti mengusulkan metode baru yang dapat digunakan untuk klasifikasi citra melalui ekstraksi fitur berupa fitur warna yang menggunakan tranformasi model warna YUV dan fitur tekstur menggunakan fungsi Gabor. Untuk klasifikasi, peneliti menggunakan Fuzzy Support Vector Machine dalam menghindari adanya daerah yang tidak dapat terklasifikasi pada metode SVM. Terdapat tiga kelas untuk klasifikasi, yaitu kelas pertanian, kelas pemukiman, dan kelas perairan. Pengujian dilakukan terhadap citra satelit dengan ukuran 256 x 256 piksel serta data latih sebanyak 450 data dengan ukuran 16 x 16 piksel. Hasil pengujian menunjukkan bahwa metode yang diusulkan peneliti dapat melakukan klasifikasi data citra dengan tingkat akurasi yang didapatkan melebihi 80%.

Kata kunci: Citra Satelit, Transformasi Citra, Fungsi Gabor, Fuzzy Support Vector Machine

Abstract

Research on the satellite image classification for grouping pixels in an image into a number of classes, so that each class can describe an entity with certain characteristics. Problems caused by errors in feature extraction or by image degradation can occur in the classification process. In this study, a new method is proposed for image classification by extracting features such as color features using a YUV color model transformation and texture features using Gabor functions. For the classification process, we use the Fuzzy Support Vector Machine to avoid unclassifiable regions in SVM method. There are three classes who used in this study, namely agricultural land, residential area, and water area. The test carried out on the satellite image size 256x256 pixels with a total number of 450 training data size 16x16 pixel image data. Tests carried out to classify images into 3 classes. Experimental results show that the proposed method is able to classify the data with an accuracy rate above 80%.

Keywords: Satellite Imagery, Image Transformation, Gabor functions, Fuzzy Support Vector Machine

Downloads

Download data is not yet available.

Referensi

ABE, SHINGO DAN TAKUYA INOUE. 2002. Fuzzy Support Vector Machine for Multiclass Problems. European Symposium on Artificial Neural Networks, Bruges, Belgia.

GONZALEZ, R.C. & WOODS E.R. 2002. Digital Image Processing. Prentice Hall. 2nd Edition.

JIANPING FAN & YAU DAVID. K. Y. 2001. Automatic Image Segmentation by Integrating Color-Edge Extraction and Seeded Region Growing. IEEE Transactions On Image Processing. Vol. 10. pp. 1454-1466.

LEI YU, ZHONGSHI HE, & QI CAO. 2010. Gabor Texture Representation Method for Face Recognition Using the Gamma and Generalized Gaussian Models. Image and Vision Computing Vol. 28. pp.177–187.

RUBERT, C., FONSECA, L., & VELHO, L. 2005. Learning Based Super-Resolution Using YUV Model for Remote Sensing Images. Proceedings of WTDCGPI.

SUPIANTO, A.A. & SUTRISNO. 2013. Transformasi Model Warna YUV dan Fuzzy Support Vector Machine untuk Klasifikasi Citra Satelit. Prosiding Seminar Nasional Teknologi Informasi dan Aplikasinya. FMIPA. Universitas Udayana, Bali.

YONGXIN G., DAN, Y., JIWEN L., BO L., & XIAOHONG Z. 2013. Active Appearance Models Using Statistical Characteristics of Gabor Based Texture Representation. J. Vis. Commun. Image R. Vol 24. pp. 627-634.

YUYONG, C. & ZENG, Z. 2009. Remote Sensing Image Classification Based on the HSI Transformation and Fuzzy Support Vector Machine. IEEE International Conference on Future Computer and Communication. pp. 632-635.

Unduhan

Diterbitkan

22-07-2015

Terbitan

Bagian

Teknologi Informasi

Cara Mengutip

Klasifikasi Citra Satelit Menggunakan Kombinasi Fitur Warna Dan Fitur Tekstur. (2015). Jurnal Teknologi Informasi Dan Ilmu Komputer, 2(2), 102-109. https://doi.org/10.25126/jtiik.201522141