14 August, 2019 | Written by: Jill Robinson
Categorized: Artificial Intelligence
Share this post:
Marwell Zoo, home to more than 140 species in a range of innovative and sustainable exhibits, is owned by Marwell Wildlife, a global conservation charity. The zoo welcomes more than half a million visitors each year and wants to encourage them to help save the planet, too.
“We all know we should recycle,” said Duncan East, Head of Sustainability at Marwell Wildlife. “But with so many different schemes in different areas it’s not always easy to know what you can and can’t recycle when out for the day. Add the fact that children are often keen to see the next animal or get to a play area, then sorting picnic waste doesn’t always get the attention it needs. We’re very excited to be working with IBM to take the stress out of recycling and make it quick, easy and fun.”
The government is also nudging business like Marwell to take action on recycling, imposing additional taxes unless they can reduce the amount of waste they send to landfill.
The students’ Wastenet prototype uses a combination of IBM technology for visual recognition, machine learning, and cloud computing. It helps visitors decide which bin to use for their waste, and provides Marwell staff with an overview of recycling across the whole site.
“You’d think throwing rubbish in a bin would be easy,” said Arthur Berkley, a Materials Science student from the University of Oxford. “But when we studied how zoo visitors used the bins, we saw that they got confused about what they can recycle. And if they throw stuff in the wrong bin, or if it’s soiled, then recycling companies can decide the whole binload is contaminated.”
“We found that out for ourselves the hard way,” said Robert Scowen, who studies Computer Science and AI at the University of Sheffield. “We spent a grim day sorting through some of the binbags and saw what a mess people made of recycling – literally.”
Their innovative waste classification system can be anyone’s “bin assistant.” Hold up an item to a scanner by the recycling bin, or the camera of your mobile phone, and an image classifier uses IBM Watson artificial intelligence to tell you which bin to put it in.
The more the prototype is used to train the system, the more the classification is refined and the greater its accuracy. The team are now taking their tech out to the zoo for testing in the field – or near the fields, at least.
As well as the tech for bins and an app for mobile phones, Wastenet combines the data in the IBM Cloud to show Marwell how visitors are using bins right across the site.
“We’ve created display dashboards to make it easy to understand,” explained Eden Sinclair, a Graphic Communication and Typography undergraduate at the University of Reading. “Marwell also want to display them in high-traffic areas around the zoo, such as cafes, to give visitors a snapshot of the impact of their waste habits.”
The students’ work at IBM’s major software development lab at Hursley, as part of “Extreme Blue,” a summer internship scheme that’s now in its twentieth year. Four different teams are given three months to design, architect, develop, and test an innovative new product for an IBM client
“Our internship is a great way to get real-world experience solving business problems,” said Kossaï Sbaï, who studies Computer Science at King’s College London. “We’re given responsibility to do research, design a product, then develop and test it with a client. And we have IBM’s experts and technology available to help us at every stage.”
This is the latest of IBM’s environmental collaborations with Marwell. A previous project used IBM ingenuity and technology to analyse data from sensors in zoo enclosures, keeping animals safe while also reducing the amount of energy used for heating.