IDENTIFIKASI PERUBAHAN TUTUPAN LAHAN BERDASARKAN KLASIFIKASI BERBASIS PIKSEL PADA FOTO UDARA RESOLUSI SPASIAL SANGAT TINGGI (STUDI KASUS : KAWASAN BOROBUDUR)

• Main Author: Mohammad Arvin, Gian
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/40430
• Institution: Institut Teknologi Bandung

Development and management of cultural sites could be done by doing an observation of land cover changes to identify the development of supporting facilities and infrastructure at certain intervals of years. Borobudur region needs to be preserved to protect and manage archeology objects for various practical-academic interests in the present and future. Pixel-based image analysis is difficult to be applied for high resolution image classification accurately and precisely because it has a lot of redundancy. Classification process limited by RGB (red green band) bands could not help classifying Borobudur region because there are different objects which has similar visual looks. Therefore, green ratio index and NDSM (Normalized Digital Surface Model) index are added to help classification process. The addition of Green Ration index aims to confirm the classification results between green and non-green objects. The addition of NDSM index aims to confirm the classification result by identifying the elevation differences of each objects. This study examined several types of classification techniques, the number of clusters to look for class objects, the number of iterations and other parameters. The utilization of Borobudur zone distribution data is used to helf classify objects in Borobudur area. The quality control of the classification results is done by analyzing the results of the table configuration matrix and object morphometry. .