Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Deo Hernando
Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Pemanfaatan Fitur Warna dan Fitur Tekstur untuk Klasifikasi Jenis Penggunaan Lahan pada Citra Drone Deo Hernando; Agus Wahyu Widodo; Candra Dewi
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 4 No 2 (2020): Februari 2020
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (548.471 KB)

Abstract

Land use is a way to use land in carrying out certain objectives. Examples of land use types are forests, rice fields, housing, roads and rivers. However, the many transfers of land use functions, such as illegal logging of land used for housing development, require the use of land use planning. Given this issue, encouraging the writing of this research to assist land use planning by classification of land use types. The author uses two image features namely color features of 8 different color spaces (CMYK, HSV, HVC, Lab, RGB, YCbCr, YIQ, and YUV) and texture features using the Support Vector Machine classification method. The data used are 25 training data and 200 test data where the amount of data for each class is the same. The tests conducted are testing the color features with the highest accuracy, testing the texture features that affect accuracy, and the combination of color and texture features with the highest accuracy. The first test result is the color feature in the HSV color space has the highest accuracy of 98%. The second test result is the accuracy of texture features affected by image size, membership distance, and angle in the GLCM calculation. The image size of 900x900 with a membership distance of 1% and a combination of 4 corner features (0o, 45o, 90o, 135o) produces the highest accuracy of 96.5%. The third test result is a combination of color features in the CMYK, HSV, HVC, Lab, YCbCr, YIQ, and YUV spaces with the texture features of the second test results yielding the highest accuracy of 99.5%.
Co-Authors Agus Wahyu Widodo Candra Dewi