DocumentCode :
3592764
Title :
An improved regional integration method based on Size-constrain Region Merging
Author :
Fengyuan Zhen ; Ling Fan ; Chao Wang
Author_Institution :
Sch. of Sci., Beijing Jiaotong Univ., Beijing, China
fYear :
2013
Firstpage :
317
Lastpage :
320
Abstract :
In this paper, an improved method is introduced to reduce the insufficiency of Size-constrain Region Merging (SCRM) segmentation method. Because the stop criterion of SCRM is all regions larger than a specific size, two regions which are both larger than the threshold value can never be merged. This will lead to over-segmentation obviously. So we propose an improved region merging method based on the result of SCRM: the merging process will continue with a similarity-constrain on the merging regions. The proposed algorithm has applied on two scenes of high resolution SAR data. The results show that our method can effectively suppress the insufficiency of SCRM method.
Keywords :
geophysical image processing; image segmentation; radar imaging; remote sensing by radar; synthetic aperture radar; SAR data; SCRM segmentation method; merging process; region merging method; regional integration method; size-constrain region merging; Algorithm design and analysis; Image resolution; Image segmentation; Merging; Remote sensing; Synthetic aperture radar; Transforms;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Synthetic Aperture Radar (APSAR), 2013 Asia-Pacific Conference on
Type :
conf
Filename :
6705078
Link To Document :
بازگشت