APPLYING CONVOLUTIONAL NEURAL NETWORKS TO IDENTIFY MOVING TARGETS IN SAR IMAGERY
Loading...
Authors
Henegar, Erik L.
Subjects
synthetic aperture radar
SAR
convolutional neural network
CNN
deep learning
DL
automatic target recognition
ATR
SAR
convolutional neural network
CNN
deep learning
DL
automatic target recognition
ATR
Advisors
Garren, David A.
Date of Issue
2021-09
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
Synthetic Aperture Radar (SAR) is a type of radar that can provide high resolution imagery regardless of time of day or weather conditions. Convolutional Neural Networks (CNNs) or other deep learning algorithms can be applied to SAR imagery to conduct Automatic Target Recognition (ATR) of high value targets. SAR is a valuable reconnaissance and surveillance capability, but it is limited in its ability to show moving targets. In SAR imagery, moving targets appear smeared, making it difficult to perform ATR. This thesis analyzed various methods for performing ATR of moving targets in SAR imagery using CNNs. Analysis was conducted through computer simulation using the Synthetic and Measured Paired and Labeled Experiment (SAMPLE) dataset to train and test the classification accuracy of a CNN algorithm. This thesis determined that out of the various analyzed methods for classifying moving targets using a CNN, the most accurate classification occurred when the CNN was trained using images of moving targets. Autofocus image processing techniques were shown to improve classification accuracy but not to acceptable levels. Future research is recommended to improve autofocus image processing techniques and to develop a method to separate stationary and moving target images for classification by CNNs trained on stationary or moving target data.
Type
Thesis
Description
Series/Report No
Department
Electrical and Computer Engineering (ECE)
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
Approved for public release. Distribution is unlimited.
Rights
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.