Surface temperature distribution in cassava planted area from Landsat 8 satellite data using object based image analysis: Case study of Mahasarakham Province, Thailand
Abstract:
This research aims to 1) classification of cassava planted, 2) classification of the surface
temperature of cassava planted, and 3) the relationship analysis between the surface temperatures
with plantations of cassava planted in Mahasarakham, Province. This study was conducted to
identify by OBIA (Object Based Image Analysis), segmentation with parameters that include Scale
parameter, Shape parameter and Compactness parameter using satellite data from LANDSAT 8
and creating vegetation index (NDVI) into the first step. And then, with the segmentation objects
has to be classified by Decision Tree Model, configuration criteria of classification into five levels
of classification. First, Root node contains data set of image segmentation and NDVI. Stage 2:
range of criteria NDIV for the classification into vegetation and non-vegetation. Stage 3: at this
levels classified from objects data set of vegetation into three stages of planted growth. Stage 4:
The criteria of the LANDSAT (Band 4) for the classify verities of planted under the Data Set of
stage 3. And the final stage of the five-level classification criteria determined by the temperature
of the thermal wave infrared (Band10).
The study found that the first step in image segmentation. Objects dataset cannot
encounter the characteristics of the landcover, there are some objects that are larger. The data set
are non-similar is in the same object. The areas identified cassava of 1,846 objects for 511,018.75
Rais or 817.63 square kilometers. This research found, the low resolution of LANDSAT can be
classify by Decision Tree model. And surface temperature can be classified criteria as good as
classification method to help identify that helping to the greater classification.