ISAR Images Segmentation Based on Spatially Variant Mixture Multiscale Autoregressive Model

ISAR images segmentation play a key role for characteristic extraction, target recognition, and target surveillance. This paper proposes a novel segmentation method of Inverse synthetic aperture radar (ISAR) images, which employees a spatially variant mixture multiscale autoregressive (SVMMAR) model...

Full description

Saved in:
Bibliographic Details
Published in:2018 IEEE 3rd Advanced Information Technology, Electronic and Automation Control Conference (IAEAC) pp. 2170 - 2174
Main Authors: Ju, Yanwei, Zhang, Yan, Guo, Feng
Format: Conference Proceeding
Language:English
Published: IEEE 01.10.2018
Subjects:
Computational modeling
Estimation
Image segmentation
Inverse synthetic aperture radar
ISAR images
least square estimation
Noise
Spatial resolution
Speckle
Stochastic processes
Surveillance
SVMMAR model
Target recognition
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ISAR images segmentation play a key role for characteristic extraction, target recognition, and target surveillance. This paper proposes a novel segmentation method of Inverse synthetic aperture radar (ISAR) images, which employees a spatially variant mixture multiscale autoregressive (SVMMAR) model to segment ISAR images. The estimation of parameters of the model is easily performed via least square estimation and expectation maximization algorithm (EM algorithm). Moreover, a kind of method for selecting number of classes at a coarser scale is proposed, which reduced computation amount greatly. In order to improve the performance, the method characterizes and exploits multiscale stochastic structure inherent in ISAR image. The advantage of our proposed segmentation approach is that it is not only fast, but also able to automatically estimate all the model parameters, and easy to implement. Therefore, the model can be exploited for ISAR automatic target recognition easily. All of that are demonstrated by the experimental results.
DOI:10.1109/IAEAC.2018.8577781