A Novel Hybrid Machine Learning-Based Model for Rockfall Source Identification in Presence of Other Landslide Types Using LiDAR and GIS
- 55 Downloads
Rockfall is a common phenomenon in mountainous and hilly areas worldwide, including Malaysia. Rockfall source identification is a challenging task in rockfall hazard assessment. The difficulty rise when the area of interest has other landslide types with nearly similar controlling factors. Therefore, this research presented and assessed a hybrid model for rockfall source identification based on the stacking ensemble model of random forest (RF), artificial neural network, Naive Bayes (NB), and logistic regression in addition to Gaussian mixture model (GMM) using high-resolution airborne laser scanning data (LiDAR). GMM was adopted to automatically compute the thresholds of slope angle for various landslide types. Chi square was utilised to rank and select the conditioning factors for each landslide type. The best fit ensemble model (RF–NB) was then used to produce probability maps, which were used to conduct rockfall source identification in combination with the reclassified slope raster based on the thresholds obtained by the GMM. Next, landslide potential area was structured to reduce the sensitivity and the noise of the model to the variations in different conditioning factors for improving its computation performance. The accuracy assessment of the developed model indicates that the model can efficiently identify probable rockfall sources with receiver operating characteristic curve accuracies of 0.945 and 0.923 on validation and training datasets, respectively. In general, the proposed hybrid model is an effective model for rockfall source identification in the presence of other landslide types with a reasonable generalisation performance.
KeywordsRockfall GIS Hybrid model LiDAR Gaussian mixture model Remote sensing
The authors acknowledge and appreciate the provision of airborne laser scanning data orthophoto images from airborne laser scanning data (LiDAR) by the Department of Planning.
This research is supported by the Centre for Advanced Modelling and Geospatial Information Systems (CAMGIS) in the University of Technology Sydney (UTS) under Grants 321740.2232335 and 321740.2232357.
- Agliardi F, Riva F, Galletti L, Zanchi A, Crosta GB (2016) Rockfall source characterization at high rock walls in complex geological settings by photogrammetry, structural analysis and DFN techniques. EGU Gen Assem Conf Abstr 18:13071Google Scholar
- Bui DT, Tuan TA, Klempe H, Pradhan Revhaug I (2016) Spatial prediction models for shallow landslide hazards: a comparative assessment of the efficacy of support vector machines, artificial neural networks, kernel logistic regression, and logistic model tree. Landslides 13(2):361–378. https://doi.org/10.1007/s10346-015-0557-6 CrossRefGoogle Scholar
- Chen W, Yan X, Zhao Z, Hong H, Bui DT, Pradhan B (2019) Spatial prediction of landslide susceptibility using data mining-based kernel logistic regression, naive Bayes and RBFNetwork models for the Long County area (China). Bull Eng Geol Environ 78(1):247–266. https://doi.org/10.1007/s10064-018-1256-z CrossRefGoogle Scholar
- Corona C, Trappmann D, Stoffel M (2013) Parameterization of rockfall source areas and magnitudes with ecological recorders: when disturbances in trees serve the calibration and validation of simulation runs. Geomorphology 202:33–42. https://doi.org/10.1016/j.geomorph.2013.02.001 CrossRefGoogle Scholar
- Dou J, Yamagishi H, Zhu Z, Yunus AP, Chen CW (2018) TXT-tool 1.081-6.1 a comparative study of the binary logistic regression (BLR) and artificial neural network (ANN) models for GIS-based spatial predicting landslides at a regional scale. In: Sassa K et al (eds) Landslide dynamics: ISDR-ICL landslide interactive teaching tools. Springer, Cham. https://doi.org/10.1007/978-3-319-57774-6_10 CrossRefGoogle Scholar
- Fanos AM, Pradhan B, Aziz AA, Jebur MN, Park HJ (2016) Assessment of multi-scenario rockfall hazard based on mechanical parameters using high-resolution airborne laser scanning data and GIS in a tropical area. Environ Earth Sci 75(15):1129. https://doi.org/10.1007/s12665-016-5936-3 CrossRefGoogle Scholar
- Giacomini A, Ferrari F, Thoeni K, Lambert C (2016) A new rapid method for rockfall energies and distances estimation. EGU Gen Assem Conf Abstr 18:5323Google Scholar
- Kotthoff L, Thornton C, Hoos HH, Hutter F, Leyton-Brown K (2017) Auto-WEKA 2.0: automatic model selection and hyperparameter optimization in WEKA. J Mach Learn Res 18(1):826–830Google Scholar
- Mote TI, Skinner MD, Taylor ML, Lyons C (2019) Site-specific rockfall risk assessments and rockfall protection structure design following the 2010/2011 Canterbury earthquake sequence. In: IAEG/AEG annual meeting proceedings, San Francisco, California, 2018, vol 5. Springer, Cham, pp 143–152Google Scholar
- Simon N, Ghani MFA, Hussin A et al (2015) Assessment of rockfall potential of limestone hills in the Kinta Valley. JSSM 10(2):24–34Google Scholar
- Varnes DJ (1984) Landslide hazard zonation: a review of principles and practice. Natural hazards, No 3. United Nations Publisher, ParisGoogle Scholar
- Wen Z, He B, Xu D, Feng Q (2016) A method for landslide susceptibility assessment integrating rough set and decision tree: a case study in Beichuan, China. In: 2016 IEEE international geoscience and remote sensing symposium (IGARSS). IEEE, pp 4952–4955. https://doi.org/10.1109/igarss.2016.7730292