Motion detection using near-simultaneous satellite acquisitions

• Published in: Remote Sensing of Environment
• Main Authors: Kääb, Andreas, Leprince, Sebastien
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2015
• Subjects: Sea ice
• Online Access: http://hdl.handle.net/10852/41895; http://urn.nb.no/URN:NBN:no-46305; https://doi.org/10.1016/j.rse.2014.08.015

A number of acquisition constellations for airborne or spaceborne optical images involve small time-lags and produce near-simultaneous images, a type of data which has thus far been little exploited to detect or quantify target motion at the Earth’s surface. These time-lag constellations were for the most part not even meant to exhibit motion tracking capabilities, or these capabilities were considered a drawback. In this contribution, we give the first systematic overview of the methods and issues involved in exploiting near-simultaneous airborne and satellite acquisitions. We first cover the category of the near-simultaneous acquisitions produced by individual stereo sensors, typically designed for topographic mapping, with a time-lag on the order of a minute. Over this time period, we demonstrate that the movement of river ice debris, sea ice floes or suspended sediments can be tracked, and we estimate the corresponding water surface velocity fields. Similarly, we assess cloud motion vector fields and vehicle trajectories. A second category of near-simultaneous acquisitions, with much smaller time-lags of at most a few seconds, is associated with along-track offsets of detector lines in the focal plane of pushbroom systems. These constellations are demonstrated here to be suitable to detect motion of fast vehicles, such as cars and airplanes, or, for instance, ocean waves. Acquisition delays are, third, also produced by other constellations such as ‘trains’ of satellites following each other and leading to time-lags of minutes to tens of minutes, which are in this contribution used to track icebergs and features of floating ice crystals on the sea, and an algae bloom. For all acquisition categories, the higher the spatial resolution of the data and the longer the time-lag, the smaller the minimum speed that can be detected.