Share Email Print

Proceedings Paper

Feature long axis size and local luminance contrast determine ship target acquisition performance: strong evidence for the TOD case
Author(s): Piet Bijl; Alexander Toet; Frank L. Kooi
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Visual images of a civilian target ship on a sea background were produced using a CAD model. The total set consisted of 264 images and included 3 different color schemes, 2 ship viewing aspects, 5 sun illumination conditions, 2 sea reflection values, 2 ship positions with respect to the horizon and 3 values of atmospheric contrast reduction. In a perception experiment, the images were presented on a display in a long darkened corridor. Observers were asked to indicate the range at which they were able to detect the ship and classify the following 5 ship elements: accommodation, funnel, hull, mast, and hat above the bridge. This resulted in a total of 1584 Target Acquisition (TA) range estimates for two observers. Next, the ship contour, ship elements and corresponding TA ranges were analyzed applying several feature size and contrast measures. Most data coincide on a contrast versus angular size plot using (1) the long axis as characteristic ship/ship feature size and (2) local Weber contrast as characteristic ship/ship feature contrast. Finally, the data were compared with a variety of visual performance functions assumed to be representative for Target Acquisition: the TOD (Triangle Orientation Discrimination), MRC (Minimum Resolvable Contrast), CTF (Contrast Threshold Function), TTP (Targeting Task Performance) metric and circular disc detection data for the unaided eye (Blackwell). The results provide strong evidence for the TOD case: both position and slope of the TOD curve match the ship detection and classification data without any free parameter. In contrast, the MRC and CTF are too steep, the TTP and disc detection curves are too shallow and all these curves need an overall scaling factor in order to coincide with the ship and ship feature recognition data.

Paper Details

Date Published: 21 October 2016
PDF: 19 pages
Proc. SPIE 9987, Electro-Optical and Infrared Systems: Technology and Applications XIII, 99870M (21 October 2016); doi: 10.1117/12.2241805
Show Author Affiliations
Piet Bijl, TNO Defence, Security and Safety (Netherlands)
Alexander Toet, TNO Defence, Security and Safety (Netherlands)
Frank L. Kooi, TNO Defence, Security and Safety (Netherlands)

Published in SPIE Proceedings Vol. 9987:
Electro-Optical and Infrared Systems: Technology and Applications XIII
David A. Huckridge; Reinhard Ebert; Stephen T. Lee, Editor(s)

© SPIE. Terms of Use
Back to Top