Two critical inputs for planning—who has access to electricity, and the location of infrastructure—are often unavailable or overly time-consuming to collect and maintain.

Building on our work to identify energy access using satellite imagery, EAP and collaborators investigated the value of high-resolution imagery collected by unmanned aerial vehicles (UAVs or drones) for identifying energy infrastructure—such as small solar panels, diesel and gasoline generators, and distribution lines. Collecting this information can support public policy and private investments in sustainable energy access. While UAV imagery has a much higher detection accuracy than satellite-based images, it introduces a number of challenges that must be overcome with advances in the machine learning algorithm.

This work explores the viability and cost-performance tradeoff of using automatic SHS detection on UAV imagery as an alternative to satellite imagery, answering the following questions: (i) what is the detection performance of SHS using drone imagery; (ii) how expensive is the drone data collection, compared to satellite imagery; and (iii) how well does drone-based SHS detection perform in real-world scenarios?

The results suggest that UAV imagery may be a viable alternative to identify very small SHS, which may make it a practical option for supporting electricity access planning strategies for achieving sustainable development goals and for monitoring the progress towards those goals. For instance, identifying the location and capacity of small PV panels would help donor agencies who require, as a precondition for companies to receive promised funds, monitoring to verify firms have installed systems. This monitoring typically requires expensive, large-scale ground surveys, but drone-based surveys could reduce the effective cost for donors and increase funds available for new connections. Similarly, developers considering entering a new market need to understand current levels of access to assess the competitive landscape, but the only option to collect data on SHS penetration now involves expensive door-to-door surveys. Providing a scalable method to measure existing access could be a game-changing development in many communities.

Team Members: Rob Fetter, Kyle Bradbury, Jordan Malof, Simiao Ren

Partners: Robert Beach and Jay Rineer, RTI International

Latest Publication: Utilizing Geospatial Data for Assessing Energy Security: Mapping Small Solar Home Systems Using Unmanned Aerial Vehicles and Deep Learning