Google Puts Its AI To Work With Dugong Conservation Effort

Google Puts Its AI To Work With Dugong Conservation Effort

Google has put its AI smarts to work, partnering with marine scientists and geoprocessing experts to create WISDAM, an image analysis software that can spot dugongs and help monitor the health of Australia’s seas.

Lead image: Amanda Hodgson.

As a marine mammal scientist, Amanda Hodgson’s core research involves monitoring marine megafauna and their critical habits. Most of that research is focused on dugongs, with the mammals being a good indicator of ecosystem health. This monitoring has long been complex since it has primarily been carried out from the air. 

Back in 2007, Amanda was one of the first in the world to investigate the potential for using drones to survey fauna. Despite the practice being a safer alternative that provided compelling data, researchers had to manually look through the thousands of images generated by the drones. Hodgson said she had dreamed of having one piece of software to integrate all the image processing needs and automate the analysis as much as possible.

“WISDAM is the outcome of many years of developing various tools for processing aerial survey imagery. My colleague, Dr Frederic Maire, developed an AI model for automating the detection of dugongs,” she said.

“Custom software was developed first by Dr Eric Kniest and then Martin Wieser to map the sightings and image footprints. Then, with the assistance of funding from the Google.org Impact Challenge, we were able to integrate these tools into WISDAM and develop many other capabilities within the software”

Using Google’s open-source TensorFlow technology, scientists can now automate image processing and save tremendous effort. WISDAM allows scientists to automatically and manually look through images, find the animals and map and assess the background environment within the pictures. In fact, WISDAM has even allowed aerial survey novices to conduct wildlife surveys. 

“There’s always been a great need to develop research tools that are accessible and affordable in developing countries that form a large proportion of the dugongs’ range. WISDAM supports drone surveys of areas that haven’t previously been surveyed and allows local research groups to conduct surveys and extract the data from those surveys to provide reliable information on animal distribution and abundance. It also allows us to obtain the benefits of imagery surveys (as opposed to surveys using real-time observers onboard piloted aircraft) and process the images within a reasonable timeframe,” said Hodgson.

Researchers, community groups and NGOs worldwide are now using WISDAM to monitor dugongs in their local areas — from the Great Barrier Reef to the Seychelles to Saudi Arabia. This is especially important in the Seychelles, as it is almost impossible to conduct surveys with piloted planes around the island nation.

“We (my colleague Dr Christophe Cleguer and myself) were able to provide remote assistance, and from this, SIF can conduct a robust, standardised dugong survey. These surveys have been very successful and are providing extremely valuable data on dugong distribution and abundance at Aldabra, as well as data on many other marine fauna species,” added Hodgson. 

Since WISDAM was first developed, its capabilities have expanded beyond dugongs, supporting surveys of turtles, whales and other marine animals. The hope is to keep developing this technology and eventually expand it to different habitats and parts of the world, meaning that, one day, anybody can help monitor wildlife and conservation efforts.

Latest News

Andie Potter On What It Means To Be A Woman In The AdTech Industry
  • Adtech

Andie Potter On What It Means To Be A Woman In The AdTech Industry

With over ten years of experience in media and marketing, Andie Potter is a passionate and experienced leader who delivers growth and innovation across digital paid and owned channels. Women Leading Tech (WLT) sat down with the Women Leading Tech award winner, Potter, to discuss how PHD incorporates technology and what it means to be […]

Why Too Few Women Are Scaling The Career Heights In The IT Sector
  • Opinion

Why Too Few Women Are Scaling The Career Heights In The IT Sector

There’s still more work to do, to ensure female candidates have a decent shot at challenging and fulfilling roles, says AVEVA director of sales, Pacific, Christine McNamara in this op-ed. Who thinks Australia is doing ok, when it comes to evening up the scales in the historically male-dominated world of IT. Women comprise 29 per […]

Opinion
Host Of New Hires At Snap
  • Social Media

Host Of New Hires At Snap

Snap Inc. has announced a host of new hires across its team in Australia, including Dina Bailey as ANZ agency lead. Lead image: L to R – Dina Bailey, Bethany Rao-Davies, Sarah Ding, Rob Fitzpatrick, Tony, Daniel King, Elise Keeling The new hires include Dina Bailey, ANZ agency lead; Daniel King, senior client partner; and […]

X Booted Out Of Social Self-Reg Code Following Inaction On Voice To Parliament Disinformation
  • Social Media

X Booted Out Of Social Self-Reg Code Following Inaction On Voice To Parliament Disinformation

X, the platform formerly known as Twitter, has been kicked out of Australia’s code for managing misinformation and disinformation online due to its lack of response to user complaints during the Voice to Parliament referendum. Lead image; Linda Yaccarino, CEO, X Twitter and subsequently X, had been a signatory to the Australian Code of Practice […]

Leonardo.Ai Accelerates Global Growth Generating 700 Million AI Images In Less Than A Year On AWS
  • Technology

Leonardo.Ai Accelerates Global Growth Generating 700 Million AI Images In Less Than A Year On AWS

Amazon Web Services (AWS) has announced that Leonardo.Ai, an Australian generative artificial intelligence (generative AI) content production platform, is creating 4.5 million new images daily on the world’s most comprehensive and broadly adopted cloud. Since launching in December 2022, Leonardo.Ai’s users have generated more than 700 million images and trained more than 400,000 custom generative […]