J. KIMS Technol Search


J. KIMS Technol > Volume 20(6); 2017 > Article
Journal of the Korea Institute of Military Science and Technology 2017;20(6):758-766.
DOI: https://doi.org/10.9766/KIMST.2017.20.6.758   
A Design of Du-CNN based on the Hybrid Machine Characters to Classify Target and Clutter in The IR Image
Juyoung Lee, Jaewan Lim, Haeun Baek, Chunho Kim, Jungsoo Park, Eunjin Koh
1The 3rd Research and Development Institute, Agency for Defense Development
2The 1st Research and Development Institute, Agency for Defense Development
적외선 영상에서의 표적과 클러터 구분을 위한 Hybrid Machine Character 기반의 Du-CNN 설계
이주영, 임재완, 백하은, 김춘호, 박정수, 고은진
1국방과학연구소 제3기술연구본부
2국방과학연구소 제1기술연구본부
In this paper, we propose a robust duality of CNN(Du-CNN) method which can classify the target and clutter in coastal environment for IR Imaging Sensor. In coastal environment, there are various clutter that have many similarities with real target due to diverse change of air temperature, water temperature, weather and season. Also, real target have various feature due to the same reason. Thus, the proposed Du-CNN method adopts human's multiple personality utilization and CNN technique to learn and classify target and clutter. This method has an advantage of the real time operation. Experimental results on sampled dataset of real infrared target and clutter demonstrate that the proposed method have better success rate to classify the target and clutter than general CNN method.
Key Words: Infrared Image, Convolutional Neural Network, Target Classification, Machine Learning, Multiple Personality
Share :
Facebook Twitter Linked In Google+ Line it
METRICS Graph View
  • 0 Crossref
  • 141 View
  • 1 Download
Related articles in J. KIMS Technol.


Browse all articles >

Editorial Office
160 Bugyuseong-daero 488beon-gil, Yuseong-gu, Daejeon 34060, Korea
Tel: +82-42-823-4603    Fax: +82-42-823-4605    E-mail: kimst@kimst.or.kr                

Copyright © 2021 by The Korea Institute of Military Science and Technology.

Developed in M2PI

Close layer
prev next