Earth Observation, Image Processing and Feature Extraction

A vast amount of digital satellite and aerial imagery is being acquired by modern Earth Observation sensors every day. The real challenge is to analyse the raw imagery quickly, extract useful, actionable information with higher accuracy, and apply it in real-world decision making and applications.

While there seems no short of image processing techniques and promises, conventional paradigms need to be critically reviewed and a mere data-driven approach is not enough. Image understanding in relation to features of interest is important. We develop accessible image analysis software tools and batch processing workflows, and make a contribution to this direction. Our latest effort is to develop dedicated software tools to rapidly process the daily fresh Copernicus/ESA Sentinel-2 and NASA/USGS Landsat-8 satellite imagery, and apply new artificial intelligence and deep learning techniques for image feature extraction.

High-resolution Digital Imagery Meets Artificial Intelligence

New advances in artificial intelligence and machine learning, especially deep learning neural networks, enable us to extract features from high-resolution satellite and aerial imagery more reliably and quickly. We apply new techniques, inject our understanding about image objects and incorporate contextual geospatial layers to classify image features with high accuracy (e.g. at 90-95%). Many factors are at play and we make our best efforts to probe deeper to achieve even better results. High-performance computing from GCP, AWS or Azure has been extensively used. Outputs are usually in GIS vector forms.

Extracting Features from High-resolution Imagery: An AI-based Processing Service

Building-level Geocoding of G-NAF (address database) for Improved Flood & Bushfire Risk Analysis in Australia

Creating Building Footprints and Terrain Features for Web and Desktop Mapping: With Examples from the NSW Coast

Mapping the Coastal Settlement in Wamberal, NSW: Two Approaches

Building-level Geocoding for Regional Flood Risk Analysis: A Case Study in Ipswich, Queensland

Monitoring Land Cover Changes at the Bushland-Urban Interface: An Image Analysis Approach

Extracted vegetation highlighted in green. (For more info, please refer to our previous project on the U.S. high-resolution vegetation mapping, link)

Spectral Exploration, Transformation and Discovery

We develop a set of image spectral analysis routines that are specific to the latest and most popular open imagery sources, including Sentinel-2 and Landsat-8 satellite imagery.

The new software Spectral Discovery provides accessible and highly efficient tools for rapid band combinations, adaptive image stretchingadvanced image pan-sharpening and exploratory image feature extraction. It has been increasingly used by geospatial professionals, environmental scientists and the general public worldwide. (International collaborators and distributors are sought to promote the software.)

Spectral Discovery software GUI (Version 03/2020)

An example of rapid image band combinations: Animated view of 336 band combinations with a Landsat-8 scene.

See Features of Interest: From Pixels to Information

We constantly ask the following important questions in image analysis: (1) What are the Features of Interest (FOI) over pixels? (2) How can they be effectively extracted from pixels in an automated digital workflow? (3) Is the classification accuracy acceptable for the applications at hand? (4) Are there any cost-effective alternatives to achieve the same objective? …

The March 2016 flooding in the U.S. South captured by Landsat-8 imagery. An example of automated classification of surface water areas (in light blue). LANDSAT_SCENE_ID = “LC80230382016080LGN00”; DATE_ACQUIRED = 2016-03-20. Link to previous blog.

See the Invisible: Applications of Short-wave Infrared (SWIR) Bands

Short-wave infrared (SWIR) bands from the latest satellites (e.g. medium-resolution Landsat-8 and Sentinel-2, and high-resolution WorldView-3) are capable of detecting unique surface features invisible to the human eye and dynamic phenomena through heavy smoke.

 The May 2016 Fort McMurray Fire (Alberta, Canada) captured by the latest Sentinel-2 satellite imagery. Left: Natural colour imagery showing full smoke; Right: False-colour SWIR imagery. Data tile ID: 12/V/VH; Date: 201605-05. Link to previous blog.

 Detecting lava flows/heat through smoke using Landsat-8 imagery. Left: Natural-colour image with RGB bands; Right: False-colour image with SWIR bands. Location: Holuhraun Lava Field, Iceland. LANDSAT_SCENE_ID = “LC82170152014249LGN00”; DATE_ACQUIRED = 2014-09-06. Images were processed by BigData Earth.

See the Detail: Advanced Image Fusion

After combining the spectral signatures of the multispectral input and the spatial sharpness of the panchromatic input, the best attributes of both inputs, the output imagery greatly assists image interpretation and visualisation.

Pan-sharpening of Landsat-8 imagery from coarse 30m-resolution (Left) to very sharp 15m-resolution (Right). Location: San Francisco; Image source: Landsat-8; LANDSAT_SCENE_ID = “LC80440342013106LGN00”; DATE_ACQUIRED = 2013-04-16. Images were processed by BigData Earth.

Applications and related blog articles:

– Monitoring Major Events with Global Earth Observation and Geospatial Big Data Analytics: 10+ New Examples (link)

– Mapping the 2019-2020 Unprecedented Bushfire Season in Australia (link 1, link 2, link 3)

– Mapping the June 2016 Sherpa Fire in Santa Barbara County, California, with Sentinel-2 Satellite Imagery (link)

– Observing the May 2016 Fort McMurray Wildfire: A Survey of Satellite Imagery Sources (including Sentinel-2) (link)

– Innovative Earth Observation Image Analysis for Wildfires and Flooding: Two U.S. Examples (link)

– Large-sized World Imagery & Elevation Basemaps (Mosaics) at 30m Resolution (link)