Thousands of cameras are hidden in the wild across the country, just waiting for an unsuspecting potoroo, dingo or feral cat to snap a selfie.
These cameras, and the massive amount of images and data they capture, are currently mostly isolated — possibly used for a research project or two, and then erased to make room on the hard drive for something else.
However, a team of Australian researchers from the University of Queensland, Sydney and Melbourne, as well as James Cook University and others, are trying to create something much more far-reaching. The team is in the early stages of building a central database, which it has dubbed the Wildlife Observatory of Australia (WildObs), to collect, store and tag this vast amount of data.
“We are creating a national standard for wildlife monitoring so researchers can, for the first time, systematically monitor wildlife across the Australian continent,” says UQ ecologist Dr. Matthew Luskin.
The idea of WildObs is to have a national “one stop shop” for researchers. Ultimately, the team hopes it will be a treasure trove of information for researchers looking for specific animals or location data, even without setting foot in the field.
However, recognizing that they are unlikely to convince researchers to give up their hard-earned data without reward, the team is developing helpful tools for ecologists in the field to make the system attractive to use.
“We want to provide tools for practitioners who set up camera traps in the field, provide tools to quickly sort images using artificial intelligence, and once those images are all sorted… we want to provide users with some nifty tools to use to set up this camera trap.” can analyze data,” UQ ecologist Zachary Amir told Cosmos.
The AI system WildObs uses is called Wildlife Insights and has already sorted over 54 million camera trap recordings from around the world.
The researchers emphasize that this is still very early. Some members of the team – like Amir – have created new datasets for the system after recently taking 130,000 photos with 25 cameras on K’gari (Fraser Island) to monitor Potoroos. Now he’s just beginning a new camera survey in Queensland’s tropical rainforests, hoping to collect camera data from nine national parks.
For him, he’s happy that the data is available as soon as possible for other researchers to analyze, although he explains that the system there will have embargoes on researchers wanting to make a first attempt at their data before going the broader one Research Foundation.
“I’m interested in species interactions and want to understand how the abundance of feral cats in one park might affect the abundance of musk-rat kangaroos in another park,” he says.
“The more people have more data, the better decisions we can all make. It’s just better for all parties.”
The WildObs team hopes to be as transparent and open as possible with their data, although there are still some issues to be resolved. For example, sensitive species could scramble their coordinates, while hunters could use their data for more nefarious purposes.
“It is still being checked whether it is not the pig hunter Joe who is using the data to find out where the wild boars are,” says Amir.
“But if you’re a researcher and you’re using the data for science, or if you’re a land steward trying to make better decisions — there shouldn’t be boundaries… We’re striving for fairly open transparency and availability of these data sources.”
