Random Forest as a Novel Machine Learning Approach to Predict Landslide Susceptibility in Kalutara District, Sri Lanka

Palliyaguru, S.T.; Liyanage, L. C.; Weerakoon, O.S.; Wimalaratne, Geethanjali

doi:10.1109/icter51097.2020.9325460

Cited by 5 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model can be exported to other file formats such as ONNX and TensorRT. ONNX is an intermediary machine learning file format used to convert between different machine learning frameworks [6]. TensorRT is a library developed by NVIDIA for optimization of machine learning model, to achieve faster inference on NVIDIA graphics processing units (GPUs) [7].…”

Section: J Export To Other File Formatsmentioning

confidence: 99%

Object Detection Driven Precision Forecasting in Curved Terrains

Surendar,

REL

et al. 2024

IJRASET

View full text Add to dashboard Cite

This study introduces an inventive approach to avalanche forecast in fastener twist locales utilizing the You Simply See Once (YOLO) protest location system. Fastener twists, with their interesting geography, posture particular challenges for avalanche forecast, requiring a custom fitted arrangement that coordinating progressed computer vision strategies. The proposed technique combines high-resolution fawning symbolism information to make nitty gritty territory models of clip twist ranges. YOLO, known for its real-time question location capabilities, is adjusted to recognize potential avalanche triggers, counting slant flimsiness and changes in vegetation cover, inside these complex scenes. Real-time observing frameworks, counting ground-based sensors and climate stations, are deliberately put to ceaselessly capture natural conditions. Integration with the YOLO-based prescient demonstrate empowers the early recognizable proof of potential avalanche dangers, encouraging the usage of focused on early caution frameworks. Community engagement remains a pivotal perspective of this approach, including neighborhood inhabitants within the improvement of departure plans and readiness techniques. The cooperative energy between YOLO-based innovation and community association makes a comprehensive arrangement for proactively overseeing avalanche dangers in clip twists

show abstract

Section: J Export To Other File Formatsmentioning

confidence: 99%

Object Detection Driven Precision Forecasting in Curved Terrains

Surendar,

REL

et al. 2024

IJRASET

View full text Add to dashboard Cite

show abstract

“…When it comes to the combination of these aspects, we can consider to be in the context of Big Data, for volume, variety, velocity, veracity, and value of data. Moreover, with the aim of producing predictions in a data driven approach, many different machine learning and deep learning algorithms have been applied in a variety of use cases: Logistic regression (LR), Support Vector Machine (SVM), Random forest (RF), Boosting, Convolutional Neural Network (CNN), as stated in [14], [17], [19], [20] [21]. The SIGMA algorithm, which was firstly developed in Emilia Romagna Region [6] and then tested in India [22], is a landslide early warning model based on the analysis of the probability related to exceedance of defined rainfall amounts.…”

Section: Introductionmentioning

confidence: 99%

Predicting and Understanding Landslide Events With Explainable AI

et al. 2022

View full text Add to dashboard Cite

Rainfall induced landslide is one of the main geological hazard in Italy and in the world. Each year it causes fatalities, casualties and economic and social losses on large populated areas. Accurate shortterm predictions of landslides can be extremely important and useful, in order to both provide local authorities with efficient prediction/early warning and increase the resilience to manage emergencies. There is an extensive literature addressing the problem of computing landslide susceptibility maps (which is a classification problem exploiting a large range of static features) and only few on actual short terms predictions (spatial and temporal). The short-term prediction models are still empirical and obtain unsatisfactory results, also in the identification of the predictors. The new aspects addressed in this paper are: (i) a short-term prediction model (1 day in advance) of landslide based on machine learning, (ii) real time features as good predictors. The introduction of explainable artificial intelligence techniques allowed to understand global and local feature relevance. In order to find the best prediction model, a number of machine learning solutions have been implemented and assessed. The models obtained overcome those of the literature. The validation has been performed in the context of the Metropolitan City of Florence, data from 2013 to 2019. The method based on XGBoost achieved best results, demonstrating that it is the most reliable and robust against false alarms. Finally, we applied explainable artificial intelligence techniques locally and globally to derive a deep understand of the predictive model's outputs and features' relevance, and relationships. The analysis allowed us to identify the best feature for short term predictions and their impact in the local cases and global prediction model. Solutions have been implemented on Snap4City.org infrastructure.

show abstract

“…Artificial intelligence methods can be divided into two types: machine learning and deep learning. Machine learning method include support vector machines [12], random forests [13], logistic regression [14], etc.…”

Section: Introductionmentioning

confidence: 99%

A Collapse Extraction Method of Remote Sensing Image Using Improved U-Net Convolution Network

2022

JRSE

View full text Add to dashboard Cite

Visual interpretation is used to identify and extract collapses on remote sensing images time-consuming and laborious. In order to solve this problem, this paper takes Liuhe County, Jilin Province, as an example, and adopts a deep learning approach to achieve automatic extraction of collapse using high-resolution remote sensing images. The classical semantic segmentation models such as DeepLabV3+, PSP-Net and U-Net built with Pytorch framework are used, and after adjusted the models and modified the hyperparameters to make them fit the collapse extraction task, a high-definition collapse remote sensing dataset has been built for training and validation. The U-Net model, which backbone is VGG-16, was found to have the best extraction effect after comparison. In order to extract the collapses in remote sensing image slices more accurately, a Pyramid U-Net model for collapse extraction is proposed by integrating the pyramid pooling module on the basis of U-Net model to enhance the multi-scale feature extraction capability of the model. The results show that the Mean Intersection over Union (MIoU) and Mean Pixel Accuracy (MPA) of thePyramid U-Net model proposed in this paper reach 86.68% and 93.11%, respectively. Compared with the traditional U-Net model, Pyramid U-Net improves the MIoU and MPA values by 0.78% and 0.15%, respectively, and further improve the effect of collapse extraction.

show abstract

Random Forest as a Novel Machine Learning Approach to Predict Landslide Susceptibility in Kalutara District, Sri Lanka

Cited by 5 publications

References 25 publications

Object Detection Driven Precision Forecasting in Curved Terrains

Object Detection Driven Precision Forecasting in Curved Terrains

Predicting and Understanding Landslide Events With Explainable AI

A Collapse Extraction Method of Remote Sensing Image Using Improved U-Net Convolution Network

Contact Info

Product

Resources

About