Digital Earth Australia for Geospatial Analysts (Course Materials)

Context

Remote sensing provides us with valuable information about the state and evolution of our environment. Processes such as soil, water, coastal erosion, crop growth, or bushfires can be monitored through the images collected by satellites. However, these collections of satellite images are often extremely large, which presents a significant challenge to users wanting to process or store these data.

Geoscience Australia has developed a platform called Digital Earth Australia (DEA), that enables users to perform large scale satellite image analysis easily on personal computers. Using a high-performance storage and computing platform provided by the National Computational Infrastructure and commercial cloud, DEA offers the tools to search and analyse high quality satellite image collections that span more than 30 years, covering the entire Australian continent.

This course is a Geoscience Australia and Australian National University, Centre for Water and Landscape Dynamics (WALD) initiative.

Goals

This course presents the tools and techniques for processing satellite imagery through practical examples.

The materials will get participants familiar with using DEA by connecting to the National Computational Infrastructure. A quick introduction to scientific Python is provided and geospatial concepts are presented using practical and interactive examples using Jupyter notebooks. By the end of the course, participants will have the necessary knowledge to use satellite data and apply it to solve a wide range of analytical problems using Python.

This training focuses on understanding and using:

• Python and Jupyter for spatial data (Introduction to scientific Python, Numpy, Matplotlib and Xarray libraries)
• DEA (examples of querying, loading, processing and exporting data)
• Geospatial analysis using DEA (examples of querying, loading, processing and exporting data)
• Machine learning to create models based on DEA data.

Next sessions:

We are currently putting together a new session for the last fortnight of November 2022. Please contact pablo.larraondo@anu.edu.au to express your interest or get further details about this course.

Accessing and running the contents

The tutorials in this course build up from basic Numpy concepts to specific datasets and analysis using the DEA infrastructure. Some of the materials in this course do not need access to the DEA Sandbox environment and can be run from this website using Binder. These tutorials are accessible in the Open Tutorials section below, to run them interactively in a Jupyter environment click on the Binder icon at the top-right corner of the notebook.

Open Tutorials

• Introduction to Jupyter notebooks

  This tutorial introduces Jupyter notebooks.

  Aug 12, 2020

• Introduction to Numpy

  This tutorial introduces numpy, a Python library for performing numerical computations in Python.

  Aug 11, 2020

• Introduction to Matplotlib for image data

  This second part introduces matplotlib, a Python library for plotting numpy arrays as images.

  Aug 10, 2020

• Introduction to remote sensing

  This tutorial introduces remote sensing and some of its underlying physical theory.

  Aug 9, 2020

• Basics of satellite image processing

  This tutorial expands the previous numpy and matplotlib functionality applied to real satellite data.

  Aug 8, 2020

• Categorical and missing data

  This tutorial explores further concepts in Numpy such as, categorical data, advanced indexing and dealing with Not-a-Number (NaN) data.

  Aug 7, 2020

• Introduction to XArray

  This tutorial introduces XArray, a Python library for working with labeled multidimensional arrays.

  Aug 6, 2020