DETEKSI PENYAKIT CERCOSPORA PADA DAUN SAGU (METROXYLON)
DENGAN EKSTRAKSI FITUR RGB BERBASIS DRONE

PDF

Abstract

MUHAMMAD RIDWAN KAMBORI.. DETECTION OF CERCOSPORA

DISEASE ON SAGO LEAVES (METROXYLON) USING DRONE-BASED RGB

FEATURE EXTRACTION (supervised by Prof. Dr. Ir. Indrabayu, ST.,MT., M.

Bus.Sys., IPM, ASEAN. Eng. and Anugrayani Bustamin, S.T., M.T.)

The Unitary State of the Republic of Indonesia (NKRI) is one of the largest

producers and suppliers of sago in the world. However, there were many crop

failures with considerable losses due to disease. One disease that generally attacks

sago is Cercospora which attacks sago plant populations starting from leaves

marked with spots and can spread quickly throughout the sago plant area.

Meanwhile, drone camera technology is currently able to detect colors with good

accuracy. In this study, a system was created that can utilize drone cameras in

detecting early attacks of Cercospora disease based on symptoms of color

changes with a height of 21 to 29 meters per meter and using image color

management on data, namely the thresholding process. The range of Hue values

used to detect disease attacks.

This study aims to determine the cluster of cercopora disease in the sago leaf

population and determine the best height of the drone in detecting cercopora

disease on sago leaves. In the research on detection of Cercospora disease on sago

leaves using the RGB. Feature extraction method Color segmentation is usually

done by converting the color space which was originally an RGB (Red, Green,

Blue) color space to an HSV (Hue, Saturation, Value) color space If the color

space is dark and there is absolutely no light, then the value of the light waves

absorbed by the eye or RGB (0,0,0) The color model is represented by the

computer program, the component values (r,g,b) Each has a color value between 0

to 255 according to the order of the respective color values Red, Green and Blue

so that the total component values are 256 when combined, there are 256 x 256 x

256 or 16.777,216 color combinations that can be formed.

Based on the results of this research, the Hue (H) range value is 0.12915 -

0.17728. to detect Cercospora disease. This detection test was carried out with

variations in height from 21 meters to 29 meters and it was found that the ideal

height value was 21 meters and the ideal height was obtained, with an accuracy of

90%.

Keyword(s): Cercospora, RGB, Color Segmentation, Range

Year : 2017

Source Project Url: Deteksi Penyakit