DocumentCode :
2916100
Title :
A new PCNN-based method for segmentation of SAR images
Author :
Li, Haiyan ; Bai, Zhengyao
Author_Institution :
Sch. of Inf. Sci. & Eng., Yunnan Univ., Kunming
fYear :
2008
fDate :
17-20 Dec. 2008
Firstpage :
1635
Lastpage :
1639
Abstract :
This study describes a new method for segmentation of Synthetic Aperture Radar (SAR) images, which integrates optimal threshold with pulse-coupled neural network (PCNN). Traditional image segmentation algorithms exhibit weak performance for SAR images due to the poor quality of SAR images. PCNN has been widely used in image segmentation. However, satisfactory results are usually obtained at the expense of time-consuming selection of PCNN parameters and the number of iterations. Simplified unit-linking PCNN with only one parameter to be determined are used in the proposed method. The method initiates segmentation with the optimal threshold so one iteration is needed. The method demonstrates accuracy and fast performance in segmentation results and in processing speed compared to those PCNN segmentation algorithms which requires determining the number of iterations and image entropy. Moreover, the method is not sensitive to noise and intensity. Experimental results show the effectiveness of the proposed method. This method aims to be possible in real-time hardware implementation.
Keywords :
entropy; image segmentation; iterative methods; neural nets; radar computing; radar imaging; synthetic aperture radar; PCNN segmentation algorithms; PCNN-based method; SAR image segmentation; image entropy; image segmentation algorithms; optimal threshold; pulse-coupled neural network; real-time hardware implementation; synthetic aperture radar images; Automatic control; Bridges; Entropy; Hardware; Image segmentation; Neural networks; Optimal control; Roads; Robotics and automation; Synthetic aperture radar; Pulse-coupled neural network; SAR image; image Segmentation; optimal threshold;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Control, Automation, Robotics and Vision, 2008. ICARCV 2008. 10th International Conference on
Conference_Location :
Hanoi
Print_ISBN :
978-1-4244-2286-9
Electronic_ISBN :
978-1-4244-2287-6
Type :
conf
DOI :
10.1109/ICARCV.2008.4795771
Filename :
4795771
Link To Document :
بازگشت