Business Challenge

The Harry Butler Institute realized that taxonomic identification of threat specimens for Barrow Island can take several days due to its remote location and needed a solution with speed and scale.

Transformation

To enhance its machine learning and deep learning capabilities, The Harry Butler Institute implemented IBM PowerAI Vision redefining the identification of invasive species.

Results

98%

confidence with assessments

Real-time

profile of biosecurity threats reduced from several days to seconds

Game-changer

for conservation and biosecurity efforts potentially across a range of species

Business challenge story

Protecting the Island

William Henry “Harry” Butler, an Australian naturalist, was known for his In the Wild with Harry Butler TV series on ABC in the 1970s. In 1970, for his efforts in the field of environmental education and conservation, Dr. William Henry Butler was awarded an Officer of the Order of the British Empire (OBE), and a decade later this title was upgraded to Commander (CBE).

In 1979 he was named joint Australian of the Year, and in 2012 he was added to the National Trust of Australia’s National Living Treasures list. Later in 2012 he was named an Officer of the Order of Australia (AO). Harry believed in educating people on how to protect the ecosystem they were part of. He was an early advocate of environmental protection and management through decades of service to Western Australia’s natural environment. His work to protect the unique environment of Barrow Island has created a worldwide benchmark for the industry. In August 2017, Murdoch University and Chevron further upheld his legacy by establishing The Harry Butler Institute. 

“With proper management, development and conservation can co-exist.” -   Dr. Harry Butler

The Harry Butler Institute is tasked with conducting ongoing surveillance to ensure that Barrow Island remains free from invasive species. The aggressive Asian House gecko, Hemidactylus frenatus, for example, is a highly mobile invasive species and time is of the essence when dealing with a potential specimen. André deSouza, Institute Business Manager at The Harry Butler Institute, explains: “Globally, managing invasive species can have a crippling effect on economies, cost billions of dollars and take several years to bring under control.” 

The whole process of identifying a single threat specimen can take several days due to the remote location of the island and time required to ship specimens to experts. Therefore, the Institute needed a fast and scalable solution that would help identify suspected biosecurity threat species.

Using artificial intelligence, the team can test an image of a suspect species and receive identification within seconds.

André deSouza, Institute Business Manager, The Harry Butler Institute

Transformation story

Identifying Invasive Species Faster

The Harry Butler Institute approached IBM, with support from Chevron, to improve managing invasive species. André comments: “Our decision to work with IBM was a natural connection. Our institute was looking for an off-the-shelf solution that would increase our chances of success. Knowing that IBM is the market leader for artificial intelligence, the Harry Butler Institute managed to find a solution with IBM where we had confidence to maintain and improve our current solution without high technical skills.”

Using IBM PowerAI Vision, which combines IBM Power System AC922 accelerated servers with deep learning toolsets and frameworks, the Institute is creating deep learning models to recognize suspected biosecurity threat species on Barrow Island. 

 

PowerAI Vision software is really easy to use and within a few months we were able to get up to 98% confidence with our assessments.

André deSouza, Institute Business Manager, The Harry Butler Institute

Results story

A Better Environment with AI Vision

Since deploying PowerAI Vision on Power System AC922 servers, triage identification of invasive gecko species has reduced from several days down to only seconds, which would enable the team to act more quickly to counter identified risks. 

André explains: “We started with conducting testing on a small dataset of gecko images and scaled up from there. PowerAI Vision software is really easy to use and within a few months we were able to get up to 98% confidence with our assessments”.

The Harry Butler Institute plans to continue to incorporate artificial intelligence in new ways to aid its efforts. There’s potential to use AI to recognize species based on their prints as with an app involving everyday citizens submitting images for species identification.

The Institute will also be looking to use AI to recognize weed species in conjunction with an autonomous robot for weed spraying. AI could also be used to check drone footage for animal tracks, such as sea turtles. Currently, environmentalists spend a lot of time manually walking paths or beaches in search of animals or animal tracks. By incorporating artificial intelligence, environmentalists can study the footage of the beach and improve the chances of spotting turtle tracks or improve the timeframe for these activities, which currently takes several days. 

“AI is practically everywhere these days, people don’t realize it.” – André deSouza

 

The Harry Butler Institute

The Harry Butler Institute champions a research space where community, business and biodiversity can co-exist; and where our efforts integrate and balance the needs and aspirations of all three sectors. Our research aims to support industry in better managing the environment. As biologists, ecologists, and conservationists we are proud to follow in the footsteps of Harry Butler.

Take the Next Step

To learn more about IBM Power System AC922, please contact your IBM representative or IBM Business Partner, or visit the following website: https://www.ibm.com/us-en/marketplace/power-systems-ac922

To learn more about IBM PowerAI Vision, please contact your IBM representative or IBM Business Partner, or visit the following website:  https://www.ibm.com/us-en/marketplace/ibm-powerai-vision