A machine learning-based computer model for the assessment of tsunami impact on built-up indices using 2A Sentinel imageries

This study aims to build a computer model to detect built-up land in the identified tsunami hazard zone based on Sentinel 2A imagery using the normalized built up area index (NBI), urban index (UI), normalize difference build-up index (NDBI), a modified built-up index (MBI), index-based builtup index (IBI) algorithms, optimized with machine learning Random Forest (RF) and extreme gradient boosting (XGboost) algorithms and the spatial patterns are predicted using the ordinary kriging (OK) method. Testing of the accuracy of the classification and optimization results was performed using the Kohen Kappa and overall accuracy functions. The results of the study show that a built-up land consisting of open land and water, settlements, industry areas, and agriculture and tourism areas can be identified using the parameters of built-up indices. The accuracy testings that were performed using overall accuracy and Kohen Kappa methods show that classification and prediction are highly accurate using XGboost machine learning, namely 91%. This study produces a novelty of finding, namely a computer model to detect and predict the spatial distribution of built-up land in 4 scales, i.e., very low, low, high, and very high based on NBI, UI, NDBI, MBI, IBI data extracted from Sentinel 2A imagery.