Share Email Print

Proceedings Paper

High Speed Photography To Provide A Three-Dimensional View
Author(s): J. S. Courtney-Pratt
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

A method is described for obtaining a 3-Dimensional record of a distant object or scene from a single viewpoint. The object or scene is illuminated by a brief flash of light, and a fast sequence of 2-Dimensional pictures recorded by the (diffusely) reflected light. In a typical case, the scene might well be a kilometer distant, the light flash might last 10-9 s and the sequence of pictures comprise, say, 100 frames at intervals of 10-9 s. The depth resolution would then be 15 cm in a total depth of 15 m and the lateral dimensions and resolutions could be similar.[1] Suppose we illuminate an object with a brief flash of light. If the flash is intense enough, we could take a photograph by the light reflected from the object. The light reflected from the nearest part of the object reaches the camera before the light from more distant points. So, if the camera has a fine enough time resolution, we could take a sequence of pictures which would record the varying time of arrival of the light, and so have a 3-dimensional record rather than the usual photographic 2-dimensional record. H. J. Caulfield and S. Somerstein[2] have suggested a somewhat similar scheme but considering only a line image across an object rather than aiming to make a full 3-D record. Several interrelated questions arise. What kind of camera? What lateral and depth resolutions? How bright a flash? Of mechanical cameras, the rotating mirror raster camera[3,4] has the finest time resolution, about 10-9 s.[5,6] Light travels 30 cm in 10-9 s, so that the depth resolution could be 15 cm. In clear still air the best angular resolution is about one second of arc. Let us consider a lateral resolution equal to the depth resolution, 15 cm. The distance R at which 15 cm subtends one second of arc is 30 km. As shown in Appendix I, the flash intensity that we would need would be about 100 joules - not at all an impossible figure with a Q-switched laser that would give a pulse of nanosecond duration - or at least a pulse with a rise-time of about a nanosecond. At shorter ranges and the same angular subtense the flash energy could be much less. Alternatively, at a range of 1 km and the same object size (15 m) and resolution (15 cm) as above the required flash energy is as before about 100 joules; and the diameter of the camera lens can be much smaller. Similar calculations can be made for dissecting image converter tube (ICT) cameras[?-10], where the time resolution can be a few picoseconds (or less)[11-13] and the corresponding depth resolution can be about a millimeter. See Appendix II. For equal lateral and depth resolution of 1 mm and with a 10 cm object the range can be 200 m. With a 20x slower sweep in the ICT, depth resolution would be 2 cm and with equal lateral resolutions of 2 cm in a 2 m object, the range could approach 4 km. Such resolutions might well be useful, particularly if one could not physically approach the object under study. The time-resolved record gives a depth resolution many times better than could be achieved at such distances with conventional coincidence-type range finders.

Paper Details

Date Published: 1 March 1983
PDF: 6 pages
Proc. SPIE 0348, 15th Intl Congress on High Speed Photography and Photonics, (1 March 1983); doi: 10.1117/12.967739
Show Author Affiliations
J. S. Courtney-Pratt, Bell Telephone Laboratories (United States)

Published in SPIE Proceedings Vol. 0348:
15th Intl Congress on High Speed Photography and Photonics
Lincoln L. Endelman, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?