Relevant spatial frequency information in the texture segmentation of night-vision imagery

Michael J. Sinai; J. Kevin DeFord; Todd J. Purkiss; Edward A. Essock

doi:10.1117/12.389335

23 June 2000 Relevant spatial frequency information in the texture segmentation of night-vision imagery

Michael J. Sinai, J. Kevin DeFord, Todd J. Purkiss, Edward A. Essock

Proceedings Volume 4023, Enhanced and Synthetic Vision 2000; (2000) https://doi.org/10.1117/12.389335
Event: AeroSense 2000, 2000, Orlando, FL, United States

Abstract

We investigated the type of spatial structure present in nighttime imagery that is perceptually relevant for human observers to be able to perform texture-based segmentation of real world scenes. Three psychophysical tasks were developed to evaluate performance of the nighttime imagery. The test imagery consisted of scenes obtained via an image-intensified low=light CCD, a long-wave infrared sensor and monochrome sensor-fusion. For one task, performance was best with the fused imagery, but for two tasks, performance with fused imagery was not improved (compared to performance with ir imagery). Spatial filtering of the scenes and further testing revealed that the mid spatial frequencies (1-4 cpd) were more critical in determining performance than either the low or high frequencies. Fourier analysis of the scenes revealed a strong relationship between power and performance, where scenes with more power (especially at the middle frequencies) supported better performance. Implications of this research are that performance depends on power at the middle frequencies for those low-level visual tasks and that fusion algorithms may be improved if this is taken under consideration.

Citation Download Citation

Michael J. Sinai, J. Kevin DeFord, Todd J. Purkiss, and Edward A. Essock "Relevant spatial frequency information in the texture segmentation of night-vision imagery", Proc. SPIE 4023, Enhanced and Synthetic Vision 2000, (23 June 2000); https://doi.org/10.1117/12.389335

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available