In infrared image, the grey distribution of background and target are instability, so it has much difficulty in the target segmentation. In this paper, a novel image segmentation algorithm is presented which is based on Contourlet transform and background complexity. Firstly, using Contourlet transform, the structure information of target and background is obtained. Next, structure similarity of target and background is computed. Finally, through the structure similarity of target and background, segmentation threshold is adjusted adaptively. If the structure similarity of target and background is low, it indicates that background is simple, segmentation threshold is set with the grey information. If the structure similarity of target and background is high, segmentation threshold is set with the structure information. The simulation experiments show that the target can be segmented truly in the complex background environment. The algorithm not only reserves the advantage of the grey segmentation in simple background environment, but overcomes the limitation of the grey segmentation in complex background environment, shows better adaptability than the traditional image segmentation methods.
[1]
Antonio F. Gómez-Skarmeta,et al.
About the use of fuzzy clustering techniques for fuzzy model identification
,
1999,
Fuzzy Sets Syst..
[2]
Robin N. Strickland,et al.
Image Sequence Processing For Target Estimation In Forward-Looking Infrared Imagery
,
1988
.
[3]
Richard G. Baraniuk,et al.
Multiscale image segmentation using wavelet-domain hidden Markov models
,
2001,
IEEE Trans. Image Process..
[4]
Guo Liang.
Image edge detection based on nonsubsampled contourlet transform
,
2011
.
[5]
Jianqi Zhang,et al.
Detection probability and detection time using clutter metrics
,
2007
.
[6]
Minh N. Do,et al.
Ieee Transactions on Image Processing the Contourlet Transform: an Efficient Directional Multiresolution Image Representation
,
2022
.
[7]
A. Bovik,et al.
A universal image quality index
,
2002,
IEEE Signal Processing Letters.
[8]
Alan C. Bovik,et al.
Mean squared error: Love it or leave it? A new look at Signal Fidelity Measures
,
2009,
IEEE Signal Processing Magazine.