Comparison and integration of machine learning and object-based algorithms for screening underlying factors and preparing a landslide classification map

Document Type : Complete scientific research article

Authors

1 Ph.D. Student in Watershed Management, Faculty of Rangeland and Watershed Management, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran

2 . Corresponding Author, Professor, Dept. of Watershed Management, Faculty of Rangeland and Watershed Management, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan

3 Professor, Dept. of Soil Science, College of Agriculture, Shiraz University, Shiraz, Iran.

4 Associate Prof., Dept. of Arid Regions Management, Faculty of Rangeland and Watershed Management, Gorgan University of Agricultural Sciences and Natural Resources, Gorgan, Iran

Abstract

Abstract
Background and objectives: Landslides, among the most destructive natural hazards following earthquakes, cause irreversible damage to both the environment and infrastructure. Due to the unique geological and climatic conditions, Iran experiences numerous landslides annually, necessitating the needs for careful studies and preventive measures. This study presents an integrated approach that leverages the capabilities of machine learning algorithms to identify effective features for landslide detection. Furthermore, it compares object-based algorithms to generate a landslide classification map utilizing Gaofen-1 satellite images. Eventually, the study includes partial dependence plot illustrating the relationship between landslides and various independent conditional factors.
Materials and methods: To identify landslides in the Mohammadabad watershed of Golestan, two Gaofen-1 satellite images from March 2023 and June 2024 were employed. Due to seasonal variations between the images, all processing was conducted separately. In the first phase, 218 landslides were identified through field visits, with 70% of these used for model training and the remaining 30% reserved for evaluating the results. The classification of satellite images and the extraction of landslide and non-landslide classes were performed using basic object-oriented classification approach involving two stages: segmentation and classification. Following image segmentation with a multi-scale segmentation algorithm, feature selection was conducted using three algorithms: neural networks, decision trees, and random forests. Factors influencing landslide identification were extracted from each segment based on the satellite images, and collinearity among features was assessed. Subsequently, these features were employed in the classification and identification of landslides using four object-based algorithms: support vector machine, nearest neighbor, decision tree, and random forest. The performance of these algorithms was compared using overall accuracy indices, Kappa coefficient, Sorensen coefficient, true positive rate, and false positive rate. Ultimately, the most effective algorithm for landslide detection using satellite imagery was determined.
Findings: The results of the feature selection analysis indicated that out of the three methods examined in this study, the random forest algorithm identified the most effective features for landslide detection using satellite images. The classification of landslides utilizing four object-oriented algorithms, support vector machine, decision tree, random forest, and nearest neighbor, revealed that the support vector machine algorithm achieved the highest accuracy of 92% and a Kappa coefficient of 0.85. This performance showed its superior capability in identifying landslides within the studied area compared to the other algorithms.
Conclusion: This study showed that the integrating of machine learning algorithms with object-based methods provides a reliable and cost-effective approach for the rapid identification of landslides using satellite images. Identifying landslides is an important step in studying this natural hazard, as the findings can offer valuable insights to managers and practitioners to enhance planning and management strategies to mitigate landslide-induced damages. For future study, utilizing high-resolution images is recommended to enable more detailed landslide identification.

Keywords

Main Subjects

References

1.Wang, L., Xiao, T., Liu, S., Zhang, W., Yang, B., & Chen, L. (2023). Quantification of model uncertainty and variability for landslide displacement prediction based on Monte Carlo simulation. Gondwana Research, 123, 27-40.
2.Heydari, N., Habibnejad, M., Kavian, A., & Pourqasmi, H. R. (2019). Landslide Susceptibility Modelling Using the Random Forest Machine Learning Algorithm in the Watershed of Rais-Ali Delvari Reservoir. Watershed research (research and development), 33 (1), 13-2. SID. https:// sid.ir/ paper/ 386906/ fa.[In Persian]
3.Heydari, N., Habibnejad, M., Kavian, A., & Pourqasmi, H. R. (2019). Landslide Susceptibility Modelling Using the Random Forest Machine Learning Algorithm in the Watershed of Rais-Ali Delvari Reservoir. Watershed research (research and development), 33(1), 126 series)), 13-2. SID. https://sid.ir/paper/ 386906/fa. [In Persian]
4.Karakas, G., Unal, E. O., Cetinkaya, S., Ozcan, N. T., Karakas, V. E., Can, R., & Kocaman, S. (2024). Analysis of landslide susceptibility prediction accuracy with an event-based inventory: The 6 February 2023 Turkiye earthquakes. Soil Dynamics and Earthquake Engineering, 178, 108491.
5.Chen, W., Xie, X., Wang, J., Pradhan, B., Hong, H., Tien Bui, D., Duan, Z., & Ma, J. (2017). A comparative study of logistic model tree, random forest, and classification and regression tree models for spatial prediction of landslide susceptibility. Catena. 151, 147-160.
6.Martha, T. R., Kerle, N., Jetten, V., van Westen, C. J., & Kumar, K. V. (2010). Characterising spectral, spatial and morphometric properties of landslides for semi-automatic detection using object-oriented methods. Geomorphology, 116, 24-36.
7.Blaschke, T., Hay, G. J., Kelly, M., Lang, S., Hofmann, P., Addink, E., & Tiede, D. (2014). Geographic object-based image analysis–towards a new paradigm. ISPRS journal of photogrammetry and remote sensing, 87, 180-191.
8.Abedini, M., Roustai, SH., & Fathi, M. H. (2016). Identification and classification of landslide types using Spectral and spatial features with an object-oriented method approach (Nasirabad to Sattarkhan Ahar Dam). Scientific-research journal of geography and planning, 22 (66), 187-205. [In Persian]
9.Fathi, M., H. Abedini, M., & Roostaei, Sh. (2019). Identification and zonation landslide prone areas using object oriented method and conditional probability theory (Bayesian theorem) Case Study: Ahar drainage basin South boundary (From Nasirabad to Sattar Khan dam). Journal of Geographic space,18 (64), 20-40. [In Persian]
10.Ghanavati, E., Ahmadabadi, A., & Gholami, M. (2019). Landslide susceptibility mapping of Kan using index of Entropy and LSM. Quantitative geomorphological researches. 8 (1), 16-33. [In Persian]
11.Kornejady, A., & Pourghasemi, H. R. (2019). Landslide susceptibility assessment using data mining models, a case study: Chehel-Chai Basin. Watershed Engineering and Management, 11 (1), 28-42. [In Persian]
12.Gasemyan, B., Abedini, M., & Roostaei, Sh. (2021). Landslide susceptibility assessment using a novel ensemble algorithm based model (Case Study: Kamyaran city, Kurdistan province). Quantitative geomorphological researches, 9 (4), 130-146. [In Persian]
13.Amatya, P., Kirschbaum, D., Stanley, T., & Tanyas, H. (2021). Landslide mapping using object-based image analysis and open source tools. Engineering Geology, 288, 106000. https://doi.org/10.1016/j.enggeo.2021.106000.
14.GoliMokhtari, L., & NaemiTabar, M. (2022). Spatial modeling and prediction of landslide risk using advanced data mining algorithms (case study: Kalt city). Quantitative geomorphology research, 10 (4), 137-116. [In Persian]
15.Sun, D., Gu, Q., Wen, H., Xu, J., Zhang, Y., Shi, S., & Zhou, X. (2023). Assessment of landslide susceptibility along mountain highways based on different machine learning algorithms and mapping units by hybrid factors screening and sample optimization. Gondwana Research, 123, 89-106.
16.Mohammad Abad Kalate watershed study project (2015). Khorasan flood control consulting engineers company. [In Persian]
17.Pourghasemi, H. R., Moradi, M., Mohammadi, B., Pradhan, R., Mostafazadeh, & Goli Jirandeh, A. (2012). Landslide hazard assessment using remote sensing data, GIS and weights-ofevidence model, South of Golestan Province, Iran. In Asia Pacific Conference on Environmental Science and Technology, Advances in Biomedical Engineering, 6, 30-36.
18.Agarwal, S., Vailshery, L., Jaganmohan, M., & Nagendra, H. (2013). Mapping Urban Tree Species Using Very High Resolution Satellite Imagery: Comparing Pixel-Based and Object-Based Approaches. ISPRS International Journal of
Geo-Information, 2 (1), 220-236. https://doi.org/10.3390/ijgi2010220.
19.Eelbode, T., Bertels, J., Berman, M., Vandermeulen, D., Maes, F., Bisschops, R., & Blaschko, M. B. (2020). Optimization for medical image segmentation: theory and practice when evaluating with dice score or jaccard index. IEEE transactions on medical imaging, 39 (11), 3679-3690.
20.Veena, V. S., Sai, S. G., Tapas, R. M., Deepak, M., & Rama, R. N. (2016). Automatic detection of landslides in object-based environment using open source tools. In Proceedings of the GEOBIA 2016, Solutions and synergies, Enschede, The Netherlands, 14-16.
21.Altarabichi, M. G., Nowaczyk, S., Pashami, S., & Sheikholharam Mashhadi, P. (2023). Fast Genetic Algorithm for feature selection-A qualitative approximation approach. In Proceedings of the companion conference on genetic and evolutionary computation. 11-12.
22.Pourgholam-Amiji, M., Ahmadaali, Kh., & Liaghat, A M. (2021). Sensitivity Analysis of Parameters Affecting the Early Cost of Drip Irrigation Systems Using Meta-Heuristic Algorithms. Iranian Journal of Irrigation and Drainage,
15 (4), 737-756. Doi: https://sid.ir/ paper/1054363/fa. [In Persian]
23.Zhang, W., Wu, C., Li, Y., Wang, L., & Samui, P. (2019). Assessment of pile drivability using random forest regression and multivariate adaptive regression splines. Georisk: Assessment and Management of Risk for Engineered Systems and Geohazards, 15 (1), 27-40.
24.R Core Team. (2024). R: A Language and Environment for Statistical Computing_R Foundation.
25.Noori, S., Nourijelyani, K., Mohammad, K., Niknam, M., Mahmoudi, M., & Andonian, L. (2012). Random forests analysis: A modern statistical method for screening in high-dimensional studies and its application in a population-based genetic association study. North Khorasan University of Medical Sciences. 3 (5), 93-101. [In Persian]
26.Vorpahl, P., Elsenbeer, H., Märker, M., & Schröder, B. (2012). How can statistical models help to determine driving factors of landslides?. Ecological Modelling, 239, 27-39.
27.Kim, J. C., Lee, S., Jung, H. S., & Lee, S. (2017). Landslide susceptibility mapping using random forest and boosted tree models in PyeongChang, Korea. Geocarto International, 33(9), 1000-1015.
28.Chang, Z., Catani, F., Huang, F., Liu, G., Meena, S. R., Huang, J., & Zhou, C. (2023). Landslide susceptibility prediction using slope unit-based machine learning models considering the heterogeneity of conditioning factors. Journal of Rock Mechanics and Geotechnical Engineering. 15 (5), 1127-1143.
29.Wang, S.C. (2003). Artificial neural network. In Interdiscriplinary computing in java programming (pp. 81-100). springer, boston, MA.
30.Liu, S., Wang, L., Zhang, W., He, Y., & Pijush, S. (2023). A comprehensive review of machine learning‐based methods in landslide susceptibility mapping. Geological Journal, 58 (6), 2283-2301.
31.Sameen, M. I., Pradhan, B., & Lee, S. (2020). Application of convolutional neural networks featuring Bayesian optimization for landslide susceptibility assessment. Catena, 186, 104249.
32.Tsangaratos, P., & Ilia, I. (2015). Landslide susceptibility mapping using a modified decision tree classifier in the Xanthi Perfection, Greece. Landslides, 13(2), 305-320.
33.Thai Pham, B., Tien Bui, D., & Prakash, I. (2018). Landslide susceptibility modelling using different advanced decision trees methods. Civil Engineering and Environmental Systems, 35(1-4), 139-157.
34.Rong, G., Alu, S., Li, K., Su, Y., Zhang, J., Zhang, Y., & Li, T. (2020). Rainfall induced landslide susceptibility mapping based on Bayesian optimized random forest and gradient boosting decision tree models-A case study of Shuicheng County, China. Water, 12 (11), 3066.
35.Grzywiński, W., Turowski, R., Naskrent, B., Jelonek, T., & Tomczak, A. (2019). The effect of season of the year on the frequency and degree of damage during commercial thinning in black alder stands in Poland. Forests, 10 (8), 668.
36.Vapnik, V. N. (1995). Introduction: Four periods in the research of the learning problem. In The nature of statistical learning theory. Springer.
37.Peng, L., Niu, R., Huang, B., Wu, X., Zhao, Y., & Ye, R. (2014). Landslide susceptibility mapping based on rough set theory and support vector machines: A case of the Three Gorges area, China. Geomorphology, 204, 287-301.
38.Tien Bui, D., Pradhan, B., Lofman, O., & Revhaug, I. (2012). Landslide susceptibility assessment in vietnam using support vector machines, decision tree, and Naive Bayes Models. Mathematical problems in Engineering, 2012 (1), 974638.
39.Ghasemian, B., Abedini, M., & Roostaei, Sh. (2016). Landslide sensitivity assessment using support vector machine algorithm (case study: Kamiyaran city, Kurdistan province), Quantitative Geomorphology Research, 6 (3), 15-36. [In Persian]
40.Chen, W., Chai, H., Zhao, Z., Wang, Q., & Hong, H. (2016). Landslide susceptibility mapping based on GIS and support vector machine models for the Qianyang County, China. Environmental Earth Sciences, 75 (6).
41.Fix, E., & Hodges, J. (1951). Discriminatory Analysis, Non Parametric Discrimination: Consistency Properties”. Technical Report 4, USA, School of Aviation Medicine Randolph Field Texas. 
42.Kotsiantis, S. B., Zaharakis, I., & Pintelas, P. (2007). Supervised machine learning: A review of classification techniques. Emerging artificial intelligence applications in computer engineering, 160 (1), 3-24.
43.Annathurai, K. S., & Angamuthu, T. (2022). Sorensen-dice similarity indexing based weighted iterative clustering for big data analytics. Int. Arab J. Inf. Technol. 19 (1), 11-22.
44.Greenwell, B. M. (2017). pdp: An R package for constructing partial dependence plots. R J. 9 (1), 421.
45.Shirvani, Z., Abdi, O., & Buchroithner, M. (2019). A synergetic analysis of Sentinel-1 and-2 for mapping historical landslides using object-oriented Random Forest in the Hyrcanian forests. Remote Sensing, 11 (19), 2300.
46.Prathom, K., & Sujitapan, C. (2024). Performance of logistic regression and support vector machine conjunction with the GIS and RS in the landslide susceptibility assessment: Case study in Nakhon Si Thammarat, southern Thailand. Journal of King Saud University-Science, 103306.
47.Sun, W., Tian, Y., Mu, X., Zhai, J., Gao, P., & Zhao, G. 2017. Loess landslide inventory map based on GF-1 satelliteimagery. Remote Sensing. 9, 314.
48.Pourghasemi, H. R. (2022). GIS, Remote Sensing, and Spatial Modeling in R. Shiraz University. Press, 254p. [In Persian]
Journal of Water and Soil Conservation
Volume 32, Issue 1
April 2025
Pages 31-55

Files

Share

How to cite

Statistics