Estimation of Sea Clutter Distribution
Parameters Using Deep Neural Network
Guoqing Wang
1 , Hao Ding
1(&) , Chaopu Wang
2 , and Ningbo Liu
3
1 Naval Aviation University, Shandong Yantai, People’s Republic of China
gqwang80@126.com, hao3431@tom.com
2 Beijing Institute of Technology, Beijing, People’s Republic of China
3 Naval Aviation University and No.23 Institute of the Second Academy,
CASIC, Shandong Yantai and Beijing, People’s Republic of China
lnb198300@163.com
Abstract. As a specific application of analytical methods on marine radar big
data, this paper introduces deep learning theory into the field of sea clutter
parameters estimation. A reasonable deep neural network model is built to
estimate the parameters of amplitude distribution models so as to overcome the
drawback of traditional methods based on statistical theory. In the proposed
method, histogram method is used to preprocess the data, then deep neural
network is trained with constructed dataset, and finally, parameter estimation
results are obtained using test dataset. Validation results with simulation data
and X-band radar-measured sea clutter data show that, compared with traditional
estimation method, the deep neural network-based estimation method can
improve parameter estimation accuracy significantly.
Keywords: Sea clutter Á Parameter estimation Á Deep neural network
1 Introduction
Sea clutter is one of the main factors that influence marine radar detection performance,
and in order to minimize its effect, accurate statistical models are required in the design
and development of advanced target detection algorithms [1–8]. Amplitude distribution
model is an important type of statistical models, and current models mainly include
Lognormal, Weibull, K, or Generalized K (GK). Traditionally, these model parameters
are estimated via the statistical theory, such as maximum likelihood estimation method,
moment-based estimation method, or numerical optimization method. Their main
feature is that the requirement for sample support is big for a single estimation, which is
not suitable for online real-time estimation. What’s more, estimation results often fall
into local optimal solutions under the condition of multiple parameters, resulting in a
decline in modeling accuracy.
In order to overcome the drawback of traditional estimation methods, a reasonable
deep neural network model is built to estimate the parameters of amplitude distribution
models. In fact, in the past two years, intelligent processing technology, which is
represented by deep learning theory, has been widely applied in radar. It is a more
efficient intelligent signal processing method with the advantages of automatic
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 326–333, 2020.
https://doi.org/10.1007/978-981-15-0187-6_38
Parameters Using Deep Neural Network
Guoqing Wang
1 , Hao Ding
1(&) , Chaopu Wang
2 , and Ningbo Liu
3
1 Naval Aviation University, Shandong Yantai, People’s Republic of China
gqwang80@126.com, hao3431@tom.com
2 Beijing Institute of Technology, Beijing, People’s Republic of China
3 Naval Aviation University and No.23 Institute of the Second Academy,
CASIC, Shandong Yantai and Beijing, People’s Republic of China
lnb198300@163.com
Abstract. As a specific application of analytical methods on marine radar big
data, this paper introduces deep learning theory into the field of sea clutter
parameters estimation. A reasonable deep neural network model is built to
estimate the parameters of amplitude distribution models so as to overcome the
drawback of traditional methods based on statistical theory. In the proposed
method, histogram method is used to preprocess the data, then deep neural
network is trained with constructed dataset, and finally, parameter estimation
results are obtained using test dataset. Validation results with simulation data
and X-band radar-measured sea clutter data show that, compared with traditional
estimation method, the deep neural network-based estimation method can
improve parameter estimation accuracy significantly.
Keywords: Sea clutter Á Parameter estimation Á Deep neural network
1 Introduction
Sea clutter is one of the main factors that influence marine radar detection performance,
and in order to minimize its effect, accurate statistical models are required in the design
and development of advanced target detection algorithms [1–8]. Amplitude distribution
model is an important type of statistical models, and current models mainly include
Lognormal, Weibull, K, or Generalized K (GK). Traditionally, these model parameters
are estimated via the statistical theory, such as maximum likelihood estimation method,
moment-based estimation method, or numerical optimization method. Their main
feature is that the requirement for sample support is big for a single estimation, which is
not suitable for online real-time estimation. What’s more, estimation results often fall
into local optimal solutions under the condition of multiple parameters, resulting in a
decline in modeling accuracy.
In order to overcome the drawback of traditional estimation methods, a reasonable
deep neural network model is built to estimate the parameters of amplitude distribution
models. In fact, in the past two years, intelligent processing technology, which is
represented by deep learning theory, has been widely applied in radar. It is a more
efficient intelligent signal processing method with the advantages of automatic
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 326–333, 2020.
https://doi.org/10.1007/978-981-15-0187-6_38