Comparing Convolutional Neural Network and Machine Learning Models in Landslide Susceptibility Mapping: A Case Study in Wenchuan County

Zhang, Sikui; Bai, Lin; Li, Yuanwei; Li, Weile; Xie, Mingli

doi:10.3389/fenvs.2022.886841

Cited by 13 publications

(7 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To sum up the results, the proposed TSDNN model is more advantageous in processing data having dynamic features. The LSM results were reclassified by the natural breakpoint method in QGIS software, and the proportion of each susceptibility class to the overall area [41], [42] was calculated (Fig. 15).…”

Section: Lsm Resultsmentioning

confidence: 99%

A Deep Neural Network Framework for Landslide Susceptibility Mapping by Considering Time-Series Rainfall

Gao,

He,

Chen

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Landslide susceptibility mapping (LSM) is of great significance in geohazard early warning and prevention. The existing LSM methods mostly used traditional static landslide conditioning factors (LCFs), which only considered the spatial features of single-pixel neighborhoods and could not extract the time-series dynamic features of developing landslides, resulting in low accuracy and insufficient reliability of LSM. To solve this problem, this study proposes to introduce time-series rainfall factors based on the traditional static factors and construct an integrated time-series dynamic neural network (TSDNN) model for LSM. A convolutional neural network adding time-distributed convolution (TDCNN) and a bidirectional long and short-term memory neural network (Bi-LSTM) are utilized to extract timeseries rainfall features, and a multiscale convolutional neural network (MSCNN) is utilized to extract static features of the static LCFs. In this study, multicollinearity analysis and GeoDetector are utilized to analyze the LCFs. Multiple evaluation metrics are utilized to analyze the proposed model performance. The results indicate that the overall accuracy has improved by introducing time-series rainfall factors, the susceptibility area of actual predicted is more refined. The study indicates that significant advantages of the proposed TSDNN model are over models (CNN, MSCNN, random forest (RF)) when processing combined static and rainfall data. This is notably evident that the accuracy is enhanced by 12.9%,10.7%, and 11.4% compared to CNN, MSCNN and RF models in the receiver operating characteristic curve (ROC) analysis, respectively. Moreover, two typical areas containing three recent landslide events validate the reliability of the proposed TSDNN model. The proposed network model framework for LSM considering time-series rainfall factors can provide new ideas and key technical support for landslide disaster prevention.

show abstract

Section: Lsm Resultsmentioning

confidence: 99%

A Deep Neural Network Framework for Landslide Susceptibility Mapping by Considering Time-Series Rainfall

Gao,

He,

Chen

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…VIF computes the extent of correlation between an independent factor and other factors in a model, and it is interpreted as VIF = 1 means factors are not correlated, VIF between 1 and 5 means factors are moderately correlated, and VIF >5 indicates high correlation among the factors. TOL is the reciprocal of VIF, and its values are between 0 and 1 (Chen and Chen, 2021;Jaydhar et al, 2022;Zhang et al, 2022). VIF and TOL are computed using the following functions: Triggering factor Rainfall Uncertain 0ཞ60, 60ཞ80, 80ཞ100, 100ཞ110, 110ཞ120, 120 mm above…”

Section: Multicollinearity Analysis Of Landslide Conditioning Factorsmentioning

confidence: 99%

“…In the past, most of the susceptibility maps were produced based on an expert's judgment which consumes a lot of time and energy, and it is difficult to quantify its accuracy due to its subjective effects. Fortunately, with the development of computer technologies such as Geographic Information Systems, Remote Sensing, and advanced data collection methods, machine learning algorithms are widely used in this field (Chowdhuri et al, 2021a;Ganga et al, 2022;Zhang et al, 2022). These developments have significantly improved LSM accuracy and efficiency.…”

Section: Introductionmentioning

confidence: 99%

Landslide susceptibility mapping using O-CURE and PAM clustering algorithms

Mwakapesa

Lan²,

Nanehkaran³

et al. 2023

Front. Environ. Sci.

View full text Add to dashboard Cite

Landslide susceptibility mapping (LSM) is a crucial step during landslide assessment and environmental management. Clustering algorithms can construct effective models for LSM. However, a random selection of important parameters, inconsideration of uncertain data, noise data, and large datasets can limit the implementation of clustering in LSM, resulting in low and unreliable performance results. Thus, to address these problems, this study proposed an optimized clustering algorithm named O-CURE, which combines: the traditional Clustering Using REpresentatives algorithm (CURE), that is, efficient for large datasets and noise data, the partition influence weight (PIW)-based method to enhance the selection of sample sets and the city block distance (CIBD) for processing of the uncertain data in CURE clustering during LSM modeling. A database containing 293 landslide location samples, 213 non-landslide samples, and 7 landslide conditioning factors was prepared for the implementation and evaluation of the method. Also, a Multicollinearity analysis was conducted to select the most appropriate factors, and all the factors were acceptable for modeling. Based on O-CURE, landslide density, and the partitioning around medoids (PAM) algorithm a susceptibility map was constructed and classified into very high (33%), high (18%), moderate (24%), low (13%), and very low (12%) landslide susceptible levels. To evaluate the performance of the O-CURE model, five statistic metrics including accuracy, sensitivity, specificity, kappa, and AUC were applied. The analysis shows that O-CURE obtained accuracy = .9368, sensitivity = .9215, specificity = .9577, kappa = .8496, and AUC = .896 is an indication of high-performance capability. Also, the proposed method was compared with the CURE algorithm, three existing clustering methods, and popular supervised learning methods. From this assessment, O-CURE outperformed the other clustering methods while showing significant and more consistent performance than the supervised learning methods. Therefore, we recommend that the O-CURE model and the constructed map can be useful in assessing landslides and contribute to sustainable land-use planning and environmental management in light of future disasters.

show abstract

“…Therefore, a landslide inventory map was constructed with 100 landslide locations identified using the combined findings of previous studies, the interpretation of Google Earth images, and fieldwork validation. According to Zhang et al (2022), there are no specific rules for choosing landslide characteristics. Thus, based on analysis of the landslide inventory map and the underlying geomorphometric conditions, eight physical characteristics, including roads and highways, stream network, DEM, slope angle, slope aspect, curvature, land use, and lithology were selected as landslide physical factors (Table 1).…”

Section: Data Acquisitionmentioning

confidence: 99%

Influencing Physical Characteristics of Landslides in Kuala Lumpur, Malaysia

Zulkafli¹,

Majid²,

Zakaria³

et al. 2023

JST

View full text Add to dashboard Cite

Landslide is one of the natural disasters that commonly occurs in terrestrial environments with slopes throughout the world. Located among countries with tropical climates, the hot and humid conditions expose Kuala Lumpur, Malaysia, to the risk of landslides. This paper aims to delineate the influencing physical characteristics of landslide occurrences in Kuala Lumpur. In this study, a 100 landslides historical data set and eight landslide factors were obtained from proper field validation and maps provided by those concerned in the government, such as distance to roads, distance to streams, elevation, slope angle, curvature, slope aspect, land use, and lithology. These factors were processed using GIS as geospatial analysis provides a useful tool for planning, disaster management, and hazard mitigation. By using ArcMap 10.8.2, a GIS software, different spatial analyses in which maps for each physical factor were layered with landslide events distribution. The weights for each factor were determined using the ANN approach resulting in the slope angle having the highest relative importance with a 100.0% value. In comparison, 8.3% represents the slope aspect as the most insignificant factor out of the eight selected characteristics for this study area. Therefore, a proper perspective and a thorough understanding of the certain slope condition have to be established for future mitigation action to support the agenda of SDG 15.

show abstract

Comparing Convolutional Neural Network and Machine Learning Models in Landslide Susceptibility Mapping: A Case Study in Wenchuan County

Cited by 13 publications

References 49 publications

A Deep Neural Network Framework for Landslide Susceptibility Mapping by Considering Time-Series Rainfall

A Deep Neural Network Framework for Landslide Susceptibility Mapping by Considering Time-Series Rainfall

Landslide susceptibility mapping using O-CURE and PAM clustering algorithms

Influencing Physical Characteristics of Landslides in Kuala Lumpur, Malaysia

Contact Info

Product

Resources

About