Open Energy Infrastructure Data Through Automated Assessment Using Satellite Imagery

Energy systems are rapidly evolving in both the developed and developing world. In many developed countries, distributed energy resources such as solar photovoltaics are experiencing significant growth while many in the developing world still lack access to modern electricity. Energy access is correlated with improvements in health, economic prosperity, education, and gender equality. Identifying optimal pathways to electrification requires detailed knowledge of local infrastructure and the precise locations of communities that would benefit most from electrification. Here, we work to fill these important information gaps by efficiently analyzing massive remotely sensed data (e.g., satellite imagery) using machine learning techniques such as convolutional neural networks. From these large raw data we extract actionable information about energy systems that can be provided to researchers and decision-makers. The goal is to develop this approach as a general tool that will enable the proliferation of open energy infrastructure data ranging from generation, to transmission, to end use consumption. To that end we present results from two specific applications of this approach: (1) using only remotely sensed data, we develop and test techniques that can be used to identify the location of distributed solar photovoltaic arrays and estimate their power capacity (in Watts) for baseline assessments and long-term policymaking and planning; and (2) we demonstrate that machine-learning algorithms applied to satellite imagery including lights at night and visible spectrum imagery provide valuable information for estimating electricity access rates over large geographic regions, at high spatial resolutions (e.g., at the village-level). These village level data may be combined with electric transmission line infrastructure locations to identify the economically and environmentally optimal pathways to electrification for rural villages via grid extension, microgrid development, or off-grid systems such as solar photovoltaics.