Technical innovation spurred by a love for photography

Passion // Project

`

It’s more than a job. These compulsively creative IBMers make innovation a way of life, creating amazing projects just for fun.

Share this page:

Technical innovation spurred by a love for photography

Andrew Trice

Technical Product Manager, IBM Cloud in Washington, D.C.

 

His passion project

AI and drone technology

Just in time for Christmas, Amazon made its first delivery by drone in rural England, boasting a click-to-deliver time of just 13 minutes. Impressive as that is, by all accounts it’s only the beginning.

In 2016, global sales of private and commercial drones hovered around USD2 billion. By 2020, PricewaterhouseCoopers LLP predicts that number will climb to USD127 billion.

But a few years ago, when the entire civil drone industry consisted of little more than a handful of intriguing toys, software developer and amateur photographer Andrew Trice was already using them as a tool to extend his vision — and seeing in them the potential for much more.

Discover what you can
do at IBM

Jobs at IBM

“I was working for Adobe at the time, and Adobe, of course, is very influential in photography circles. As a photographer, I began using drones to do landscape photography, to get panoramas and vistas you just can’t get from the ground. You want to fly 50 feet out over the water to get a different perspective on the sunrise? There’s no other way to get it. But it always seemed to me that drones had so much more potential than just taking pictures.”

Targeting developers with drone technology

Trice joined IBM as a developer advocate in 2014 — a role for which the enthusiastic innovator was perfectly suited. It was his job to draw the attention of developers to IBM services like Bluemix and entice them to develop apps that incorporate those services.

“The drone thing started out as a hobby for me, but I found it was an effective way to get developers focused on IBM services. Everybody loves drones.”

By 2020, the U.S. drone market is expected to climb to USD127 billion.

Trice, now a technical product manager in IBM Cloud, combined his love for drones with emerging AI tools and created the Skylink drone application. “I was able to use a lot of the services we have on Bluemix, plus Watson Visual Recognition to enable image analysis — the contents of pictures or video — while the drone is still in flight.”

More passion projects

Smarter traffic managment

Trice began developing the idea in early 2016. “I did the first part of it as a demo for the InterConnect conference last February. I literally wrote the code in one night that enables the drone to upload images to the cloud and use Bluemix. Then, after InterConnect, it sat on the back burner for a long time until I had some free cycles. I spent maybe a week refining it, then did a blog post and a video about it.”

Just a placeholder image

Trice on IBM DeveloperWorksTV explaining hail damage insurance analytics provided by drones and IBM Watson Visual Recognition.

A role for drones in the insurance industry

That combination of AI and drone technology got a lot of attention, both in and out of IBM. The app was featured in an article on TechCrunch, a news site focused on technological innovation.

“This started as a passion project, but it morphed into something I was spending a lot of my work time on,” he says.

Around the same time that Trice’s drone project was generating all this interest, Dr. Rubina Ohanian, of the IBM Financial Services Center of Competency group, was putting the finishing touches on an idea to pitch cognitive drone technology to insurance clients. She knew the idea had enormous potential, but needed someone with development and drone expertise to turn the concept into a working application.

With an end-to-end solution like this, the insurance company can say within minutes, ‘X percent of your roof is damaged and this is the cost analysis to repair it’.

“Dr. Ohanian asked if I could help develop a proof of concept showing how a cognitive drone could identify and assess rooftop hail damage so her team could take it to their customers and build engagements with it,” says Trice. “In return, I said my team needed to be able to talk about it publicly so we could continue to generate developer interest.

“We decided to demo the app at both Gartner Symposium and World of Watson in October, and at that point it became my whole focus for about six weeks,” he continues. “I was working on it all day, then coming home and working on it at night until midnight. It was crazy, but we developed it successfully to the point where we were comfortable showing it at World of Watson.”

Demo explores a better way to inspect roof damage

The demo featured a use case relevant to the USD1.2 trillion U.S. insurance industry. “The drone can perform complex tasks in ways that are safer and more efficient than we could do them in the past,” says Trice. “Using images taken by the drone, Watson can recognize incidents of hail damage on shingle roofs. Every time a large hailstone hits a roof, it leaves an indentation. Shingle roofs are basically granules stuck to tar, so when something hits the surface, it leaves a mark.”

The standard insurance industry method of assessing hail damage is for adjusters to climb the roof and take pictures of damages—a process that is both time-consuming and potentially dangerous.

“From end to end, the whole process can take weeks or months,” explains Ohanian. “You need to get an adjuster to climb up on the roof, take pictures, send the pictures to the insurance company to review and approve the damages, then send it to a third party for repair estimates—a lengthy process for insurers and their customers.

At first, people tend to think of them as toys, but once they get a vision of what drones can do, they run with it.

“With this solution, the adjuster pulls up in the driveway, sends the drone up, takes images of the roof and uploads the images to our cloud, from which point the process will be automated. The images go into Bluemix; Watson Visual Recognition analyzes the images; and our partner, DataWing, creates a 3D reconstruction of the home and provides roof measurements. With an end-to-end solution like this, the insurance company can say within minutes, ‘X percent of your roof is damaged and this is the cost analysis to repair it.’ The goal is for the adjuster to write a check on the spot, expediting the claim process while reducing costs and improving efficiency.”

Developers see an expansive future for cognitive drone apps

The drone market is set to rapidly infiltrate myriad industries and public services, with innovators like Trice leading the way. “They’re starting to be used in everything from inspection of infrastructure to crop analytics, bridge inspections, surveying… the drone is a flying sensor. It captures information that can be used to drive decisions that impact real-world things.”

Among possible uses for drones equipped with visual recognition:

  • Firefighters can attach thermal imaging cameras to a drone and immediately see where the hotspots are in a forest fire.
  • Search and rescue teams can use them to look for heat signatures to find someone lost in the wilderness.
  • Farmers can use them to determine the health of crops and make decisions affecting crop yield.
  • Surveyors can create a 3D visualization of a construction site as the structure is being built.

Using drones to keep people out of harm’s way

“One of my favorite use cases for drones is in power line inspection,” says Trice. “To do this job, people literally hang out of the window of a helicopter flying at low altitude and take pictures. It’s one of the most dangerous jobs in the country, and people die doing it every year. If drones are used to gather the imagery, these inspections can be performed without putting people in harm’s way.

“The drone market is being compared to where cell phones were 15 years ago,” he adds. Commercial drones themselves range in price from under USD1,000 to upward of USD100,000, with most falling in the low end of that scale. “At first, people tend to think of them as toys, but once they get a vision of what drones can do, they run with it. We’re still at the very early stages of developing solutions, but the potential is enormous.”

Share this page: