Utilization of GIS Machine Learning for Mapping of Building Shapes in Tibang Village, Banda Aceh City, Aceh, Indonesia

Tibang is one of the villages in Banda Aceh which is located in a high tsunami-hazard zone. Today, many initial aid houses in the area have grown larger than the original, and new housing complex has begun to emerge. Therefore, it is necessary to find an efficient method to map the buildings in the village, so that an up-to-date buildings map of the village can be prepared quickly. Machine learning GIS is the use of machine learning algorithms in processing and analyzing spatially-based GIS data. This study aimed at applying the Support Vector Machines (SVM) method of GIS Machine learning for classifying buildings in a very-high-resolution image of Tibang Village, SVM has been used with Object-Based Image analysis (OBIA) to classify UAV-based images for various Land-Use / Land Cover (LULC) applications. The stages in classifying buildings using SVM are starting with processing high-resolution data from UAV photos to create an ortho-mosaic image. The image was then segmented using OBIA techniques and then was classified using the SVM method. The result of the SVM classification was 230 house objects identified in the village of Tibang, Syiah Kuala District with the number of training data samples used for the building class being 29 with various variations of the shape of the building objects identified in the UAV images.