Wildlife Protection Solutions (WPS) sits in the nexus between the most remote places in the world and burgeoning animal conservation science. It provides the monitoring technology that conservation groups need to keep watch over wild places and protect wildlife. Conservationists use remote cameras to gather image data about the status of the species they protect, but the number of images that must be analyzed before action can be taken is overwhelming. WPS overcomes this barrier in collaboration with Microsoft AI for Earth, supported by Azure technologies for a species-preserving solution.
“I know that Microsoft is a teammate that’s going to be around in the long term. We continue to rely on and appreciate its support of the conservation community.”
Eric Schmidt, Executive Director, Wildlife Protection Solutions
Myriad dangers face wildlife in our twenty-first century world: poachers, changing climate, and shrinking, degraded habitats that increasingly bring endangered species into conflict with humans. But conservation groups are getting an edge with technology. Wildlife Protection Solutions (WPS) gives conservation a much-needed advantage with the unimaginably difficult task of monitoring vulnerable animal populations all over the planet by using Microsoft AI for Earth tools to analyze vast amounts of image data. Azure high-performance computing (HPC) underpins the nonprofit’s wpsWatch solution, which offers hope for animals, people, and the planet.
Protecting species with pixels
Conservationists are passionate about saving species, but the race to protect wild animals is grounded in a holistic approach. “Beyond the wonder and joy that biodiversity brings to everyone, we know that everything is connected,” says Matt Hron, Director of Product and Customer Success at Wildlife Protection Solutions. “Protecting wild species helps protect this big network of interrelated living and non-living environment that’s part of the world where we all live.”
Putting enough boots on the ground is one of the most difficult challenges that conservationists face. Conservation teams first tackled the problem with remote cameras in critical habitats, sending crews to manually place the devices and revisit them at regular intervals to replace batteries and exchange secure digital (SD) cards. While the imagery that was gathered was valuable for understanding trends, it couldn’t address illegal activity as it occurred.
Enter the current generation of remote cameras, which transmit images over the Global System for Mobile Communications (GSM) cellular, wireless, and satellite networks. Solar technology powers cameras in many areas, relieving human crews of replacing batteries. All these technology advances go a long way to sustaining an electronic monitoring system, but without a dedicated team to interpret the enormous volume of images that are surging into the database, their usefulness was curtailed. The answer lies in AI technology.
In 2019, WPS engaged with the Microsoft AI for Earth team to contribute images to MegaDetector, an AI model developed by the AI for Earth team to accelerate the processing of camera trap images. It’s a mutually beneficial relationship—WPS uses MegaDetector to help sharpen and refine its wpsWatch monitoring solution, and the infusion of images it supplies contributes to the constant improvement of the model. For Wildlife Protection Solutions cofounder and Executive Director Eric Schmidt, the wins achieved to date are sweet. “I love the fact that our team comes to work every day and builds a tool that protects wildlife,” he says. “We have a direct hand in preventing poaching.”
Supporting the image lifeline with Microsoft Azure
When a remote camera fires, image data is emailed from the remote cameras via the SendGrid service and is then parsed with the WPS API. The photograph is stored using Microsoft Azure Blob Storage, and the metadata is consigned to WPS by Microsoft SQL Server. The photo forwards to various AI image recognition solutions to determine what is in the photo, such as vehicles, people, or animal species of interest, and then it’s delivered to the relevant team in the field. WPS also maintains a private web instance that’s hosted on Azure and runs on an ASP.NET Angular web application. “We take full advantage of Azure services,” says Matt Morrissette, Director of Technology at Wildlife Protection Solutions.
For him, the processing power of Azure HPC makes a critical difference. Because HPC combines compute, storage, and networking solutions into a unified suite that connects to the workload orchestration services underlying an HPC application, it combines price and performance gains. “We’ve noticed that using MegaDetector improved threat detection accuracy,” he says. “We also found that it processes images faster than other AI models we’ve tried—in some cases, by 50 percent.” This speed comes from the use of Microsoft Azure Virtual Machines that include NVIDIA V100 GPUs (graphics processing units)—wpsWatch sends images to an API that runs the MegaDetector model on a cluster of GPUs, providing sub-second response time.
The Microsoft Azure Stack is key to optimizing the work of the small WPS team. “We save a lot of time because we’re not chasing down ways to connect disparate technologies,” says Morrissette. “The streamlined nature of the Azure Stack brings a lot of benefit to a small team like ours, and as time goes on, that’s more and more apparent.”
Gaining on the problem in unexpected ways
A host of unforeseen benefits continues to unfold with wpsWatch. “In the past weeks, we’ve caught copper thieves in some reserves,” says Hron. “But beyond detecting wildlife threats, we can also help local people have better interactions with wildlife for the betterment of both. When we detect an elephant or lion close to a community, for example, we can get a real-time notification to those people so they can be prepared to chase off the animal and protect their livestock.”
Humans aren’t the only risk to wildlife populations. “MegaDetector provides broad support for use cases in any location where we look for people or animals,” says Hron. “That’s why it’s so useful for detecting invasive species, such as cats, dogs, and goats.” Adds Morrissette, “We noticed a wildfire developing thanks to images from one of our cameras. The local group was able to immediately respond and put it out before it became a major inferno.”
Other organizations use MegaDetector for a different problem. Conservation biologists studying wildlife populations also use motion-triggered cameras, then often end up sorting through millions of images that contain no animals at all—leaves blowing in the wind, or cars driving by on a park access road. Microsoft makes MegaDetector available for these scenarios through a different API that uses Microsoft Azure Batch to process tens of millions of images, eliminating those not relevant to wildlife surveys. This gives ecologists more time to focus on conservation planning rather than image sorting.
Magnifying the benefits: non-governmental organizations collaborating with Microsoft
WPS understands that changing minds and attitudes is also key to successful conservation efforts. “We believe that one of the greatest ways to create empathy for the wild world is to show it to people,” says Schmidt. “Placing 360 degree virtual reality (VR) cameras in protected areas for up to four days can provide a very unique insight into the lives of wild animals.” His WPS team developed a VR app with which people can tour the globe, focusing on various partner-protected locations to visit specific animal habitats, like a water hole or a migration path.
And the unintended interactions can be even richer. The team has seen lions pick cameras up and carry them for what Schmidt calls a “lion’s-eye view,” drawing viewers even closer into a fascinating realm that few humans have ever seen. “We’re slowly expanding to enhanced gamification, more education, and conservation information. Eventually, we’ll add a donate function so that people can support the work of our partner organizations,” he adds. “We use Azure to host the entire application, and to optimize handling an intense amount of data.”
“Working with Microsoft has been fantastic,” says Schmidt. “MegaDetector has been key to scaling the wpsWatch platform, which now hosts more than 100 sites across almost 20 countries, and to supporting the number of inbound images that we get in a scalable, highly reliable way.” He also appreciates the help Microsoft provides for infrastructure costs, which helps the nonprofit maximize its investment in other ways. “Microsoft has provided Azure usage credit for wpsWatch, in addition to the educational outreach program we’re building with virtual reality, another data-intensive innovation.”
Schmidt values not just the innovation, but the reliability of a committed technology company. “As I plan for our future, I know that Microsoft is a teammate that’s going to be around in the long term,” he concludes. “We continue to rely on and appreciate its support of the conservation community.”
Find out more about Wildlife Protection Solutions on Twitter, Facebook, and LinkedIn.
“We save a lot of time because we’re not chasing down ways to connect disparate technologies. The streamlined nature of the Microsoft stack brings a lot of benefit to a small team like ours, and as time goes on, that’s more and more apparent.”
Matt Morrissette, Director of Technology, Wildlife Protection Solutions
Follow Microsoft