Improving GIS-Based Landslide Susceptibility Assessments with Multi-temporal Remote Sensing and Machine Learning

Abstract

This study developed a systematic approach with machine learning (ML) to apply the satellite remote sensing images, geographic information system (GIS) datasets, and spatial analysis for multi-temporal and event-based landslide susceptibility assessments at a regional scale. Random forests (RF) algorithm, one of the ML-based methods, was selected to construct the landslide susceptibility models. Different ratios of landslide and non-landslide samples were considered in the experiments. This study also employed a cost-sensitive analysis to adjust the decision boundary of the developed RF models with unbalanced sample ratios to improve the prediction results. Two strategies were investigated for model verification, namely space- and time-robustness. The space-robustness verification was designed for separating samples into training and examining data based on a single event or the same dataset. The time-robustness verification was designed for predicting subsequent landslide events by constructing a landslide susceptibility model based on a specific event or period. A total of 14 GIS-based landslide-related factors were used and derived from the spatial analyses. The developed landslide susceptibility models were tested in a watershed region in northern Taiwan with a landslide inventory of changes detected through multi-temporal satellite images and verified through field investigation. To further examine the developed models, the landslide susceptibility distributions of true occurrence samples and the generated landslide susceptibility maps were compared. The experiments demonstrated that the proposed method can provide more reasonable results, and the accuracies were found to be higher than 93% and 75% in most cases for space- and time-robustness verifications, respectively. In addition, the mapping results revealed that the multi-temporal models did not seem to be affected by the sample ratios included in the analyses.

Susceptibility to Gully Erosion: Applying Random Forest (RF) and Frequency Ratio (FR) Approaches to a Small Catchment in Ethiopia

“…The RF model is a multivariate nonparametric machine learning technique developed by Breiman [91]. The RF is a powerful decision tree classifier that predicts well when there is missing data, avoids over-fitting problems, produces more stable results, and is less sensitive to multicollinearity than other machine learning algorithms (e.g., support vector machine (SVM) and classification and regression tree (CART)) [30,88,92,93]. It is also known for predicting gully erosion very well compared to other machine learning algorithms [41].…”

Review on remote sensing methods for landslide detection using machine and deep learning

“…Above enhanced versions are not only used for landslide susceptibility analysis but also used for landslide displacement prediction, 20 selection of relevant conditioning factors, 20 landslide area 20 . According to Reference 35, the capability of SVM, DT (Decision Tree), Neural Network (NN), RF (Random Forest) with GIS datasets and remote sensing images has been explored for landslide susceptibility mapping. RF has received expanded attention in recent years because of the following advantages: exquisite accuracy, highest processing speed, potential to investigate high‐dimensional data.…”

“…Accuracy of object‐based ML technique is completely based on segmentation quality 34 ; therefore, more research on improving segmentation quality is expected to be done. According to Reference 35, Random forest is one of the most frequently used method for landslide susceptibility analysis. Reference 36 suggested landslide mapping results can be improved by using CNN.…”

GIS-Based Landslide Susceptibility Mapping for Land Use Planning and Risk Assessment

“…In the recent years, Geographical Information System (GIS) technology has been largely used for landslide susceptibility assessing and mapping, frequently combined with data detected by innovative techniques, e.g., satellite remote sensing and light detection and ranging (LiDAR) images. GIS-based models allow to manage big volumes of data, both in terms of file size and geographical scale, and to perform a dynamic and on-going landslide susceptibility zonation, which represents an essential requirement for a proper land-planning and risk mitigation [16][17][18][19].…”