Multi-Source Data Fusion Based on Ensemble Learning for Rapid Building Damage Mapping during the 2018 Sulawesi Earthquake and Tsunami in Palu, Indonesia

Abstract

This work presents a detailed analysis of building damage recognition, employing multi-source data fusion and ensemble learning algorithms for rapid damage mapping tasks. A damage classification framework is introduced and tested to categorize the building damage following the recent 2018 Sulawesi earthquake and tsunami. Three robust ensemble learning classifiers were investigated for recognizing building damage from Synthetic Aperture Radar (SAR) and optical remote sensing datasets and their derived features. The contribution of each feature dataset was also explored, considering different combinations of sensors as well as their temporal information. SAR scenes acquired by the ALOS-2 PALSAR-2 and Sentinel-1 sensors were used. The optical Sentinel-2 and PlanetScope sensors were also included in this study. A non-local filter in the preprocessing phase was used to enhance the SAR features. Our results demonstrated that the canonical correlation forests classifier performs better in comparison to the other classifiers. In the data fusion analysis, Digital Elevation Model (DEM)- and SAR-derived features contributed the most in the overall damage classification. Our proposed mapping framework successfully classifies four levels of building damage (with overall accuracy >90%, average accuracy >67%). The proposed framework learned the damage patterns from a limited available human-interpreted building damage annotation and expands this information to map a larger affected area. This process including pre- and post-processing phases were completed in about 3 h after acquiring all raw datasets.

Source Characteristics of the 28 September 2018 Mw 7.4 Palu, Indonesia, Earthquake Derived from the Advanced Land Observation Satellite 2 Data

“…C-band (wavelength of 5.6 cm) Sentinel-1 SAR data were provided freely by the European Space Agency. These post-event data have been used for rapid building damage mapping for this event [23]. However, due to the short wavelength, the Sentinel-1 SAR coseismic interferograms show strong decorrelation in this area, so they were not used in this study.…”

Drawback in the Change Detection Approach: False Detection during the 2018 Western Japan Floods

“…An important constraint is the availability of a pre-event image. To the best of our experience, a temporal baseline from few months to about a year performs fairly well to identify damages in urban areas [12][13][14][15][16][17][18]. However, other types of land use may exhibit constant and systematic changes through time.…”

Tsunami Damage Detection with Remote Sensing: A Review

“…The evaluation was made in the use of a random forest [55] and two further modifications termed as rotation forest [97] and canonical rotation forest [98]. The classifiers were used to identify the building damage during the 2018 Sulawesi, Indonesia earthquake tsunami.…”

“…The damage map provided by the Copernicus Emergency Management Service were used as training data [99]. The proposed framework learned the damage patterns from a limited available human-interpreted building damage annotation and expands this information to map a larger affected area [55]. Table 3 summarized the studies reported in this chapter.…”

“…The structure of affected buildings represent complex patterns, especially, the affected buildings are surrounded by debris and water [46], which increase the difficulty of damage characterization and detection. In addition, the more accurate damage detection requires observation from multi-sensors [55,117]. The resulting data are complex, with a multi-spatial resolution, spectral and temporal acquisition requiring innovative approaches.…”