kushanavbhuyan

Building-Footprint-Extraction

This repository is for the Python Elective course at the Faculty of ITC, University of Twente, Netherlands. The repository consists of the data, model and instructions required to perform building footprint extraction from satellite imagery using a U-Net model.

Building-Identification-for-Exp-Vul-Risk-Assessment

Information about buildings is, sufficed to say, a very important aspect not just for urban land registry or transportation but also for disaster/hazard risk assessment. Specifically, typological attributes of buildings like number of residents living in them, number of floors, and many more. The study aims at figuring and capturing the typological attributes of the buildings by incorporating deep learning and other proxy information as a means of detecting and characterising the buildings.

Delineating-failure-kinematics

This repository contains code for our upcoming work on landslide kinematic separation, meaning, the demarcation between the source and runout zones.

Forest-vegetation-loss

This is a GEE code for forest monitoring using Sentinel-1 to assess vegetation loss for the western Ugandan cityof Kasese.

HR-GLDD-A-Global-Landslide-Mapping-Data-Repository

Landsifier

A python library to estimate likely triggers of mapped landslides

Landslide-mapping-on-SAR-data-by-Attention-U-Net

Large-scale-multi-spatiotemporal-landslide-mapping

By using the pre-trained models, this method enables quick and simple mapping of landslides at various spatiotemporal scales. The method also offers the adaptability of re-training a pretrained model to identify landslides caused by both rainfall and earthquakes on different target locations.

Multi-Temporal-Landslide-Mapping-Nepal

Introducing a transfer learning approach to map landslides temporally over a given spatial location.

Uncovering-landslide-failure-types

This repository contains sample codes for the automatic detection of landslide movement types based on topological information, using a landslide's 3D shape.

kushanavbhuyan/Large-scale-multi-spatiotemporal-landslide-mapping

By using the pre-trained models, this method enables quick and simple mapping of landslides at various spatiotemporal scales. The method also offers the adaptability of re-training a pretrained model to identify landslides caused by both rainfall and earthquakes on different target locations.

kushanavbhuyan/HR-GLDD-A-Global-Landslide-Mapping-Data-Repository

kushanavbhuyan/Uncovering-landslide-failure-types

This repository contains sample codes for the automatic detection of landslide movement types based on topological information, using a landslide's 3D shape.

kushanavbhuyan/Building-Footprint-Extraction

This repository is for the Python Elective course at the Faculty of ITC, University of Twente, Netherlands. The repository consists of the data, model and instructions required to perform building footprint extraction from satellite imagery using a U-Net model.

kushanavbhuyan/Multi-Temporal-Landslide-Mapping-Nepal

Introducing a transfer learning approach to map landslides temporally over a given spatial location.

kushanavbhuyan/Building-Identification-for-Exp-Vul-Risk-Assessment

Information about buildings is, sufficed to say, a very important aspect not just for urban land registry or transportation but also for disaster/hazard risk assessment. Specifically, typological attributes of buildings like number of residents living in them, number of floors, and many more. The study aims at figuring and capturing the typological attributes of the buildings by incorporating deep learning and other proxy information as a means of detecting and characterising the buildings.

kushanavbhuyan/Landslide-mapping-on-SAR-data-by-Attention-U-Net

kushanavbhuyan/Forest-vegetation-loss

This is a GEE code for forest monitoring using Sentinel-1 to assess vegetation loss for the western Ugandan cityof Kasese.

kushanavbhuyan/Landsifier

A python library to estimate likely triggers of mapped landslides

kushanavbhuyan/Delineating-failure-kinematics

This repository contains code for our upcoming work on landslide kinematic separation, meaning, the demarcation between the source and runout zones.

kushanavbhuyan/Flood-extent-and-exposure-analysis

The application of Google Earth Engine is used for analysis of flood extent and damage assessment of western Ugandan town of Kasese.

kushanavbhuyan/gem-global-active-faults

A global homogenised database of active faults maintained by the GEM Foundation

kushanavbhuyan/Landslide-mapping-with-Fully-Convolutional-Networks

An FCN based model for detecting landslide footprints based on existing landslide inventories of Kerala, 2018.

kushanavbhuyan/Landslide_shape_tools_Taylor_2018

Contains code for measuring landslide ellipticity and length-to-width ratio

kushanavbhuyan/Physically-Based-Hazard-Modelling

A physically-based multi-hazard modelling approach with OpenLisem and PC Raster.

kushanavbhuyan/TransUnet

This repo reproduces the results of TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation

kushanavbhuyan/TransUNet-tf

Tensorflow Implementation of TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation