The Usage of the Multi Task Learning Concept in Landslide recognition with Artificial Neural Nets (original) (raw)
Related papers
Identification of Landslide Areas with Neural Nets for Hazard Analysis
Within the framework of risk analysis the knowledge of the existence and location of hazard zones is an essential prerequisite. Besides the well-known deterministic and statistical methods, the use of artificial neural nets is a new and very promising approach for automated hazard analysis. The advantage provided by neural nets is their ability to handle non-linearities as well as model and parameter uncertainties. Several different types of nets have been developed for landslide recognition and different training methods have been tested. One interesting approach is to use neural nets not only for the main task, the landslide recognition, but also to solve smaller problems or tasks associated to the main problem individually. Therefore neural nets for scarp recognition were developed, whose output is used as an additional input for the nets for landslide recognition. In contrast to teams of neural nets, where the nets work parallel, the nets are here switched in line. It is of particular interest that the neural nets can identify different types of mass movements with one and the same net and also classify the run out area in one step as a hazard zone. In tests carried out to identify landslides in test areas in the Eastern Alps the best nets have classified up to 86% of the areas correctly.
Artificial Neural Networks applied to landslide susceptibility assessment
Geomorphology, 2005
Landslide hazard mapping is often performed through the identification and analysis of hillslope instability factors, usually managed as thematic data within geographic information systems (GIS). In heuristic approaches, these factors are rated by the attribution of scores based on the assumed role played by each of them in controlling the development of a sliding process. Other more refined methods, based on the principle that the present and the past are keys to the future, have also been developed, thus allowing less subjective analyses in which landslide susceptibility is assessed by statistical relationships between past landslide events and hillslope instability factors. The objective of this research is to define a method with the ability to forecast landslide susceptibility through the application of Artificial Neural Networks (ANNs). The Riomaggiore catchment, a subwatershed of the Reno River basin located in the Northern Apennines (Italy), was chosen as an ideal test site, as it is representative of many of the geomorphological settings within this region.
intechopen.com
In last few years, a new approach to landslide hazard evaluation using GIS, data mining using fuzzy logic, and artificial neural network, neuro-fuzzy models have been applied (investigated the landslide susceptibility in Malaysia. Pradhan & Lee (2010a) evaluated three models for landslide susceptibility analysis using frequency ratio, logistic regression and artificial neural network model. Pradhan & Lee (2010a) analyzed the rainfall precipitation in the Penang area using back-propagation neural networks. However, they could not have a detail landslide hazard analysis due to lack of rainfall intensity data. Slope stability and rainfall intensity is very important factors causing most of the landslides in Malaysia. Besides these two important factors of rainfall and slope, soil weight and distance to drainage are also important factors in some regions. Pradhan et al., (2009) investigated the landslide susceptibility using fuzzy model at Penang Island and they pointed out some important factors, such as topographic slope, topographic aspect, topographic curvature, distance to drainage, lithology, distance to faults, soil texture, landcover, vegetation index and accumulated rainfall intensity. The objective and motivation of this study is to demonstrate artificial neural network model with five different training strategies for landslide susceptibility mapping with the aid of GIS. In order to get a stable and reliable result, in this paper, nine geological and geomorphological factors including, topographic slope, topographic aspect, topographic curvature, stream power index (spi), distance from drainage;, flow length, flow accumulation, topographic wetness index, distance to road, lithology, distance to the fault lines, soil types, land cover, and ndvi were used to predict landslide susceptible areas. These fourteen factors constructed an ANN using the back propagation algorithm for landslide susceptibility mapping. To meet the objectives, firstly the ANN model was trained using training sites which can be directly utilized for the landslide susceptibility analysis as long as the recorded nine factors are fed into an ANN model. Five different training samples were selected to train the ANN in order to avoid bias effect in the final results. Finally, the results of the landslide susceptibility maps were validated using the existing landslide location data with the aid of receiver operating characteristics (ROC) approaches.
Journal of Hydroinformatics, 2014
ABSTRACT Susceptibility assessment of areas prone to landsliding remains one of the most useful approaches in landslide hazard analysis. The key point of such analysis is the correlation between the physical phenomenon and its triggering factors based on past observations. Many methods have been developed in the scientific literature to capture and model this correlation, usually within a GIS framework. Among these, the use of neural networks, in particular the multi-layer perceptron (MLP) networks, has provided successful results. A successful application of the MLP method to a basin area requires the definition of different model strategies, such as the sample selection for the training phase or the design of the network structure. The present study investigates the effects of these strategies on the development of landslide susceptibility maps by applying different model configurations to a small basin located in northeastern Sicily (Italy), where a number of historical slope failure events have been documented over the years. Model performances and their comparison are evaluated using specific metrics.
2014
Landslides are one of the most widespread natural hazards that cause damage to both property and life every year, and therefore, the spatial distribution of the landslide susceptibility is necessary for planning future developmental activities. In this paper the artificial neural network (ANN) technique is tested for developing a landslide susceptibility map in Turbolo River catchment, North Calabria, South Italy. Landslides were mapped through air-photo interpretation and field surveys, by identifying both the landslide depletion zones (DZs) and accumulation zones (AZs); and relevant geo-environmental thematic layers pertaining to landslide predisposing factors were generated using air-photo interpretation, field surveys and Geographic Information System (GIS) tools. Ten predisposing factors were related to the occurrence of landslide: lithology, faults, land use, drainage network, and a series of topographic factors: elevation, slope, aspect, plan curvature, to-pographic wetness index (TWI) and stream power index (SPI). In order to evaluate and validate landslide susceptibility, the DZs were divided in two groups using a random partition strategy. The first group (training set) was used to prepare the susceptibility map, employing a back-propagation learning algorithm in the Idrisi Taiga software. The second group (testing set) was used to validate the landslide susceptibility model, using the confusion matrix and the receiver operating characteristic (ROC) curve. The susceptibility map was classified into five susceptibility classes: very low, low, moderate, high, and very high. About 46% of the study area falls in high to very high susceptible classes and most of the DZs mapped (87.3%) occur in the same classes. The validation results showed satisfactory agreement between the susceptibility map and the DZs locations; over 85% of the DZs of the validation set are correctly classified falling in high and very high susceptibility areas. Also, the ROC curve had shown an area under curve (AUC) value of 0.90 which demonstrates the robustness and good reliability of the landslide susceptibility model. According to these results, we conclude that the map produced by the artificial neural network is reliable and the methodology applied in the study produced high performance, and satisfactory results, which may be useful for land planning policy.
Neural Network Aided Evaluation of Landslide Susceptibility in Southern Italy
International Journal of Modern Physics C, 2011
Landslide hazard mapping is often performed through the identification and analysis of hillslope instability factors. In heuristic approaches, these factors are rated by the attribution of scores based on the assumed role played by each of them in controlling the development of a sliding process. The objective of this research is to forecast landslide susceptibility through the application of Artificial Neural Networks. In particular, given the availability of past events data, we mainly focused on the Calabria region (Italy). Vectors of eight hillslope factors (features) were considered for each considered event in this area (lithology, permeability, slope angle, vegetation cover in terms of type and density, land use, yearly rainfall and yearly temperature range). We collected 106 vectors and each one was labeled with its landslide susceptibility, which is assumed to be the output variable. Subsequently a set of these labeled vectors (examples) was used to train an artificial neural network belonging to the category of Multi-Layer Perceptron (MLP) to evaluate landslide susceptibility. Then the neural network predictions were verified on the vectors not used in the training (validation set), i.e. in previously unseen locations. The comparison between the expected output and the artificial neural network output showed satisfactory results, reporting a prediction discrepancy of less than 4.3%. This is an encouraging preliminary approach towards a systematic introduction of artificial neural network in landslide hazard assessment and mapping in the considered area.
Remote Sensing
Natural hazards generate disasters and huge losses in several aspects, with landslides being one of the natural risks that have caused great impacts worldwide. The aim of this research was to explore a method based on machine learning to evaluate susceptibility to rotational landslides in an area near Cuenca city, Ecuador, which has a high incidence of these phenomena, mainly due to its environmental conditions, and in which, however, such studies are scarce. The implemented method consisted of an artificial neural network multilayer perceptron (ANN MLP), generated with the neuralnet R package, with which, by means of different backpropagation algorithms (RPROP+, RPROP−, SLR, SAG, and Backprop), five landslide susceptibility maps (LSMs) were generated for the study area. A landslide inventory updated to 2019 and 10 conditioning factors, mainly topographical, geological, land cover, and hydrological, were considered. The results obtained, which were validated through the AUC-ROC valu...
Neural networks and landslide susceptibility: a case study of the urban area of Potenza
Natural Hazards, 2008
For those working in the field of landslide prevention, the estimation of hazard levels and the consequent production of thematic maps are principal objectives. They are achieved through careful analytical studies of the characteristics of landslide prone areas, thus, providing useful information regarding possible future phenomena. Such maps represent a fundamental step in the drawing up of adequate measures of landslide hazard mitigation. However, for a complete estimation of landslide hazard, meant as the degree of probability that a landslide occurs in a given area, within a given space of time, detailed and uniformly distributed data regarding their incidence and causes are required. This information, while obtainable through laborious historical research, is usually partial, incomplete and uneven, and hence, unsatisfactory for zoning on a regional scale. In order to carry this out effectively, the utilization of spatial estimation of the relative levels of landslide hazard in the various areas was considered opportune. These areas were classified according to their levels of proneness to landslide activity without taking recurrence periods into account. Various techniques were developed in order to obtain upheaval numerical estimates. The method used in this study, which was applied in the area of Potenza, is based on techniques derived from artificial intelligence (Artificial Neural Network—ANN). This method requires the definition of appropriate thematic layers, which parameterize the area under study. These are recognized by means of specific analyses in a functional relationship to the event itself. The parameters adopted are: slope gradient, slope aspect, topographical index, topographical shape, elevation, land use and lithology.
Lecture Notes in Computer Science, 2013
Landslides are significant natural hazards in many areas of the world. Mapping the areas that are susceptible to landslides is essential for a wise territorial approach and should become a standard tool to support land-use management. A landslide susceptibility map indicates landslide-prone areas by considering the predisposing factors of slope failures in the past. In the presented work, we evaluate the landslide susceptibility of the urban area of Senise and San Costantino Albanese towns (Basilicata, southern Italy) using an Artificial Neural Network (ANN). In order, this method has required the definition of appropriate thematic layers, which parameterize the area under study. To evaluate and validate landslide susceptibility, the landslides have been randomly divided into two groups, each representing the 50% of the total area subject to instability. The results of this research show that most of the investigated area is characterized by a high landslide hazard.
Environmental Monitoring and Assessment, 2020
Landslide susceptibility maps can be developed with artificial neural networks (ANNs). In order to train our ANNs, a digital elevation model (DEM) and a scar map of one previous event were used. Eleven attributes are generated, possibly containing redundant information. Our base model is formed by, essentially, one input (the DEM), eleven attributes, 30 neurons, and one output (susceptibility). Principal components (PCs) group information in the first projected variables, the last ones can be expendable. In the present paper, four groups of models were trained: one with eleven attributes generated from the DEM; one with 8 out of 11 attributes, in which 3 were eliminated by their high correlation with others; other, with the data projected over its PCs; and another, using 8 out of 11 PCs. The used number of neurons in hidden layer is 30, calibrated based on a complexity analysis that is an in-house developed method. The ANN models trained with the original data generated better statistical results than their counterparts trained