Publication detail

Infrared image segmentation using growing immune field and clone threshold

YU, X. ZHOU, Z. GAO, Q. LI, D. ŘÍHA, K.

Original Title

Infrared image segmentation using growing immune field and clone threshold

Type

journal article in Web of Science

Language

English

Original Abstract

Fast and accurate segmentation of infrared target is the basis of automatic target recognition, but there is a problem that it is easy to appear the significant differences of target areas in segmentation. In order to solve this problem, in this paper a new method based on growing immune field and clone threshold for segmentation of infrared targets is introduced. First, according to the global gray information, obtain the best threshold of the image using the clonal selection algorithm for global threshold segmentation. And the seed region is selected based on global threshold segmentation. Second, the source seeds are obtained by comparing the similarity threshold with seed region. Third, the growing immune field is adjusted automatically for region growing through the source seeds. Finally, the segmented image is obtained by immune region growing. The simulation results show that the target information gained by the proposed method is complete and exact. This resultgreatly facilitates the target recognition.

Keywords

Infrared image Clonal selection algorithm Region growing Growing immune field

Authors

YU, X.; ZHOU, Z.; GAO, Q.; LI, D.; ŘÍHA, K.

Released

9. 1. 2018

Publisher

Elsevier

ISBN

1350-4495

Periodical

INFRARED PHYSICS & TECHNOLOGY

Year of study

88

Number

2018

State

Kingdom of the Netherlands

Pages from

184

Pages to

193

Pages count

10

BibTex

@article{BUT148564,
  author="Xiao {Yu} and Zijie {Zhou} and Qiang {Gao} and Dahua {Li} and Kamil {Říha}",
  title="Infrared image segmentation using growing immune field and clone threshold",
  journal="INFRARED PHYSICS & TECHNOLOGY",
  year="2018",
  volume="88",
  number="2018",
  pages="184--193",
  doi="10.1016/j.infrared.2017.11.029",
  issn="1350-4495"
}