Cover Image

TEKNIK PENGOLAHAN CITRA DIGITAL UNTUK PENDUGAAN KADAR AIR TANAH LAHAN KERING DENGAN PENANGKAP CITRA INFRARED WEBCAM

Susanto Budi Sulistyo, Masrukhi , Ardiansyah , Afik Hardanto

Abstract


Moisture content is one of important soil characteristics in agricultural systems and is
often needed to determine other soil characteristics. Gravimetric method is widely used to
measure soil moisture content. Although the result is quite accurate, however it is very time
consuming. In this research, soil moisture content of arid field has been analyzed by image
processing method using webcam and infrared webcam as the image-captured devices. This
research aimed to estimate soil moisture content from images captured by webcam and infrared
webcam and to define the most appropriate visual properties as the soil moisture content
prediction parameter. Soil samples used in this research was from arid field of Gunung Tugel
Village Banyumas Regency. Thirty samples of arid field soils were collected using soil ring
samplers in 5 cm of soil depth. All samples were wetted until they reached a saturated
condition. After that, the samples were stored for 30 days under room temperature condition
until the soils became dry. Everyday the weight of each soil was measured and the images were
captured by using webcam and infrared webcam. The images were then extracted by using
image processing to obtain visual parameters such as RGB index as well as HSI colour and
texture feature. Each of the visual parameters was then determined its correlation with soil
moisture content and the R2 of the correlations. The final step of the analysis was calculating
the mean percentage error (MPE) of the moisture content estimation. A visual parameter which
has the least MPE would be used as the prediction parameter. The results of this research
showed that soil moisture content of arid field could be predicted by using visual parameters
from images captured by webcam and infrared webcam. Intensity of the soil images captured by
infrared webcam resulted prediction error lower than that by webcam.

Full Text:

View Journal

Refbacks

  • There are currently no refbacks.