Disaster Risk Monitoring Using Satellite Imagery United Nations Satellite Centre UNOSAT
Type
Course
Location
Web based
Date
Self-paced, open-enrolment event
Is this event associated with a learning outcome?
Yes
Does the event include an objective assessment of learning?
No
Duration of event
1 Days
Programme Area
Satellite Imagery and Analysis
Specific Target Audience
No
Website
https://courses.nvidia.com/courses/course-v1:DLI+S-ES-01+V1/?ncid=so-link-599935#cid=dli03_so-link_en-us
Price
$0
Event Focal Point Email
edoardo.nemni@unitar.org
Registration method
Public – by registration
Mode of delivery
E-learning
Languages
English
Background

Learn how to build and deploy a deep learning model to automate the detection of flood events using satellite imagery. This workflow can be applied to lower the cost, improve efficiency, and significantly enhance the effectiveness of various natural disaster management use cases.

 

THIS COURSE IS CO-DEVELOPED BY UNOSAT and NVIDIA. IN ORDER TO REGISTER FOR THIS COURSE, PLEASE KINDLY ENROLL ON THIS WEBSITE: https://courses.nvidia.com/courses/course-v1:DLI+S-ES-01+V1/?ncid=so-link-599935#cid=dli03_so-link_en-us

Learning objectives

By participating in this is course, you will learn how to:

  • Implement a machine learning workflow for disaster management solutions
  • Use hardware accelerated tools to process large satellite imagery data
  • Apply transfer-learning to cost-efficiently build deep learning segmentation models
  • Deploy deep learning models for near real-time analysis
  • Utilize deep learning-based model inference to detect and respond to flood events

 

Additional information

Prerequisites:

  • Competency in the Python 3 programming language
  • Basic understanding of Machine Learning and Deep Learning concepts (specifically variations of CNNs) and pipelines
  • Interest in understanding how to manipulate satellite imagery using modern methods

Tools, libraries, frameworks used: NVIDIA DALI, NVIDIA TAO Toolkit, NVIDIA TensorRT, and NVIDIA Triton Inference Server