Using computer vision to help utilities prevent wildfires

2-minute read

As a developer, I always appreciate having the chance to do work that can make a real difference in the world—that can help save lives, prevent hardship, or help the environment. In my current role, I have the chance to contribute a project that meets all three of these ideals by using computer vision and machine learning to help utility companies prevent wildfires.  

In 2020 alone, California wildfires burned more than 4 million acres of land, destroyed more than 10,000 structures, and caused over $10 billion in damages. The state’s utility providers are well aware of their role in fire prevention and many, like our client, are turning to artificial intelligence to support the effort.

Our client came to us for a solution that will help them become aware of and address problems with power lines and related equipment that could create fire hazards. The company has tens of thousands of these assets spread across a wide service area and has been using drones to collect images that enable them to identify potential issues. With their drones collecting millions of images each year, they cannot rely solely on human employees to assess the visual data.

My team is working with the client on using machine learning models—specifically, computer vision—to augment and eventually automate the asset evaluation and risk identification process. We’re also building a platform that will deploy those models and run them at scale on millions of images quickly and cost-effectively.

In my role I serve as the tech lead for two teams—a data science team working on the models and a machine learning engineering team working on building the platform—which gives me the chance to oversee development, provide a roadmap, and coach my fellow team members. It’s a pleasure to collaborate with such talented individuals on a daily basis, and the support we receive from our project managers has been instrumental to our collective success.

Working on this project gives me the opportunity to work with advanced methods in computer vision and cutting-edge technologies in machine learning engineering, as well as large-scale data sets. Thanks to our efforts, our client has already saved millions of dollars by reducing inspection times. We’ve also had the chance to see our work gain high-profile exposure and attract attention from the tech and utility industries: the project has received a CIO award for technology innovation, and our work has been highlighted in major tech news outlets, including the Wall Street Journal.

Even more importantly, we all come to work each day knowing that what we do makes a real difference in the lives of our neighbors, our communities, and our planet.