4 ms·
A slightly more in-depth PR about this program is also available on the DARPA web site: https://www.darpa.mil/news-events/2015-05-22 https://www.darpa.mil/news
by crdrost 8y ago
A slightly more in-depth PR about this program is also available on the DARPA web site:
https://www.darpa.mil/news-events/2015-05-22 https://www.darpa.mil/news-events/2015-05-22
Presumably "active light" means "we flash some light source, possibly invisible to the naked eye, and see what comes back" and this enables some of that transit-time type of statistics that they are trying to collect? So assuming one color at a time, you would get a time-series of images, a sort of 3D model of your normally-2D image. By collecting both polarizations separately and flashing a polarized light source you would get two independent 3D models of the light's propagation through a room.
- the8472 8y agoIt seems to be a basic research program asking the question "what if we capture light properties in other ways than time-aggregated photon intensities on a pixel raster behind some color filter" and build both theory and lab experiments to find answers to that question. So they're casting a wide net that could cover polarization, transit-time, hyperspectral imaging or phase information.
- theoh 8y agoSee these multiple independent MIT research results (which happen to be mentioned in an Atlantic article linked to by another commenter): http://people.csail.mit.edu/klbouman/cornercameras.html http://people.csail.mit.edu/klbouman/cornercameras.html http://web.media.mit.edu/~raskar/trillionfps/ http://web.media.mit.edu/~raskar/trillionfps/ http://openaccess.thecvf.com/content_cvpr_2018/papers/Baradad_Inferring_Light_Fields_CVPR_2018_paper.pdf http://openaccess.thecvf.com/content_cvpr_2018/papers/Barada... http://people.csail.mit.edu/torralba/research/accidentalcameras/ http://people.csail.mit.edu/torralba/research/accidentalcame...