On-Orbit Demonstrations of Proactive Tasking of Glint Imagery

The phenomenon of glint occurs when an observer catches specular reflections of sunlight from the surface of open water. An imaging spectrometer can measure crucial information about local atmospheric composition by observing absorption spectra from glint between 0.3 and 2.4 microns. The genesis of the Carbon Mapper mission came from a need to use high-quality hyperspectral data to locate methane point source emitters at the facility scale to support mitigation action. As part of its overall strategy for global greenhouse gas monitoring, the Carbon Mapper mission is designed to utilize glint imagery to study offshore emissions of CH
4
and CO
2
. The partnership includes the non-profit Carbon Mapper, Planet, and JPL. The Carbon Mapper mission is designed to utilize glint imagery to study off-shore emissions of greenhouse gasses. In most satellite sensing applications, glint is often a serendipitous event. It is usually captured by coincidence when the satellites are in the right configuration at the right time, or even skipped to avoid sensor saturation. Having a dedicated glint imagery product requires a reliable methodology of tasking a satellite to autonomously capture glint images. This paper presents novel approaches taken to address the above-mentioned problem, which were then validated by tasking Planet's existing fleet of satellites and are planned for the upcoming Tanager satellites which are fulfilling the Carbon Mapper mission. Specifically, we present a formalized methodology to predict future glint windows over a specific region. We then study various tasking approaches that describe the satellite's actions during these windows to autonomously acquire glint captures. These actions are then demonstrated by orbiting satellites, and their captures are then analyzed. Tasking an imaging satellite requires precise window prediction models of imaging opportunities. Collecting a glint image, however, also requires the target on the ground to act as a perfect mirror during the imaging event. This is modeled as additional constraints on the opportunity generation model: (1) the Sun-satellite relative azimuth is required to be 180 degrees, and (2) the satellite elevation must equal the Sun elevation. This model is used to find opportunities to capture glint over desired targets. Satellites from Planet's two operational constellations, SkySats, and Doves, are tasked for validation. SkySats and Doves operate on different tasking philosophies, so we test two different tasking philosophies on these constellations. SkySats employ a “Target Track” approach wherein the satellite camera is pointed at the desired target as the satellite orbits over the target. The Doves, on the other hand, employ a “Pushbroom” approach wherein the satellite maintains a fixed, off-axis attitude as it passes over the target region. The two strategies were deployed on these constellations and were able to demonstrate successful glint captures. While both strategies can validate our window predictions, the relative longevity of the satisfaction of specular reflective constraints offered by the Pushbroom strategy, demonstrated on Doves, offers a favorable advantage and is therefore considered as a nominal glint capture strategy for the Carbon Mapper mission.