Slop failure model based on UAV images using GWR technique / Nor Ziana Mohd Zain

Slope failure is the serious natural phenomena occur in Malaysia especially during the monsoon season. The majority of catastrophes in this country have occurred in hillside areas, resulting deaths, injuries, and property destruction. Over the past few decades, numerous research has created landslide susceptibility maps using a variety of methods. This study was focused on monitoring a slope failure using Unmanned Aerial Vehicles (UAV) images. The primary data that were used in the study were images data from UAV in Lojing Gua Musang. Orthomosaic, Digital Surface Model and Digital Terrain Model were constructed through photogrammetry processes. A Digital Elevation Model was also created, from which the slope, elevation and aspect were created using the algorithm in ArcGIS software. Geographically Weighted Regression (GWR) provide better prediction for local estimation where four related factors were calculated and extracted from the spatial database and used to analyse slope failure model. Compared with the global logistic regression model, the Akaike Information Criteria was improved by 1409.565, the adjusted R-squared was improved from 0.244 to 0.995, and the sum of square ANOVA of this analysis was improved from 37331.061 to 223.206. The comparisons obtained from the models show that geographically weighted regression has higher predictive performance. As the result, the appropriate combination colours and symbols of the slope failure map was produced for spatial prediction model in order to be applied for slope failure monitoring approaches by the authorities and environmentalist.