The OSIRIS‐REx Asteroid Sample Return Mission is the third mission in National Aeronautics and Space Administration (NASA)'s New Frontiers Program and is the first U.S. mission to return samples from an asteroid to Earth. The most important decision ahead of the OSIRIS‐REx team is the selection of a prime sample‐site on the surface of asteroid (101955) Bennu. Mission success hinges on identifying a site that is safe and has regolith that can readily be ingested by the spacecraft's sampling mechanism. To inform this mission‐critical decision, the surface of Bennu is mapped using the OSIRIS‐REx Camera Suite and the images are used to develop several foundational data products. Acquiring the necessary inputs to these data products requires observational strategies that are defined specifically to overcome the challenges associated with mapping a small irregular body. We present these strategies in the context of assessing candidate sample sites at Bennu according to a framework of decisions regarding the relative safety, sampleability, and scientific value across the asteroid's surface. To create data products that aid these assessments, we describe the best practices developed by the OSIRIS‐REx team for image‐based mapping of irregular small bodies. We emphasize the importance of using 3‐D shape models and the ability to work in body‐fixed rectangular coordinates when dealing with planetary surfaces that cannot be uniquely addressed by body‐fixed latitude and longitude.
This paper presents the photogrammetric foundations upon which the Community Sensor Model specification depends, describes common coordinate system and reference frame transformations that support conversion between image sensor (charge‐coupled device) coordinates to some arbitrary body coordinate, and describes the U.S. Geological Survey Astrogeology Community Sensor Model implementation (https://github.com/USGS-Astrogeology/usgscsm). We present a new image support data specification that provides the position, pointing, timing, and metadata information necessary to properly locate a pixel or observations location on a body and describe a system architecture designed to explicitly identify the responsibilities of software components within a larger pipeline or analytical environment. This paper concludes with a set of experiments that illustrate positional and pointing error in the sensor location and the impact on the computed surface location.