Utilization of Deep Learning for Mapping Land Use Change Base on Geographic Information System: A Case Study of Liquefaction

Abstract

This study aims to extract buildings and roads and determine the extent of changes before and after the liquefaction disaster. The research method used is automatic extraction. The data used are Google Earth images for 2017 and 2018. The data analysis technique uses the Deep Learning Geography Information System. The results showed that the extraction results of the built-up area were 23.61 ha and the undeveloped area was 147.53 ha. The total length of the road before the liquefaction disaster occurred was 35.50 km. The extraction result after the liquefaction disaster was that the area built up was 1.20 ha, while the buildings lost due to the disaster were 22.41 ha. The total road length prior to the liquefaction disaster was 35.50 km, only 11.20 km of roads were lost, 24.30 km. Deep Learning in Geographic Information Systems (GIS) is proliferating and has many advantages in all aspects of life, including technology, geography, health, education, social life, and disasters.