Share Email Print

Proceedings Paper

Multiscale differential fractal feature with application to target detection
Author(s): Zelin Shi; Ying Wei; Shabai Huang
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

A multiscale differential fractal feature of an image is proposed and a small target detection method from complex nature clutter is presented. Considering the speciality that the fractal features of man-made objects change much more violently than that of nature's when the scale is varied, fractal features at multiple scales used for distinguishing man-made target from nature clutter should have more advantages over standard fractal dimensions. Multiscale differential fractal dimensions are deduced from typical fractal model and standard covering-blanket method is improved and used to estimate multiscale fractal dimensions. A multiscale differential fractal feature is defined as the variation of fractal dimensions between two scales at a rational scale range. It can stand out the fractal feature of man-made object from natural clutters much better than the fractal dimension by standard covering-blanket method. Meanwhile, the calculation and the storage amount are reduced greatly, they are 4/M and 2/M that of the standard covering-blanket method respectively (M is scale). In the image of multiscale differential fractal feature, local gray histogram statistical method is used for target detection. Experiment results indicate that this method is suitable for both kinds background of land and sea. It also can be appropriate in both kinds of infrared and TV images, and can detect small targets from a single frame correctly. This method is with high speed and is easy to be implemented.

Paper Details

Date Published: 27 July 2004
PDF: 8 pages
Proc. SPIE 5430, Acquisition, Tracking, and Pointing XVIII, (27 July 2004); doi: 10.1117/12.541985
Show Author Affiliations
Zelin Shi, Shenyang Institute of Automation (China)
Ying Wei, Northeastern Univ. (China)
Shabai Huang, Shenyang Institute of Automation (China)

Published in SPIE Proceedings Vol. 5430:
Acquisition, Tracking, and Pointing XVIII
Michael K. Masten; Larry A. Stockum, Editor(s)

© SPIE. Terms of Use
Back to Top