Application of SENTINEL-2 and LANDSAT 8 to classifying second-level land use classification in the SathingPra peninsular, Songkhla province

Abstract

This research is data application of SENTINEL-2 and LANDSAT 8 to classifying second-level land use classification in Songkhla Lake Basin by remote sensing technology. That the corrective comparison between a second-level land use classification between an Unsupervised Classification by twenty types of land use categorizing was founded the SENTINEL-2 providing accuracy higher than LANDSAT 8 which the accuracy in summer was shown 44.94% and 38.38% respectively and Supervised Classification, by analyzing the reliability of the data compared to the field checkpoint data is 86.18% - 94.62% via the SENTINEL-2 as well. The six types of correct land use by classifying are urban and built-up, perennial plants, rotational farming, deciduous forest, beach forest, and beach in turn. While, a horticulture is not suitable for the land use classification which the least accuracy is 25%. The land use classification in Songkhla Lake Basin contributed agricultural areas have the most land use, urban and built-up, water source areas, forest areas and miscellaneous areas in order which is consistent the land use data of the Land Development Department (LDD).