Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks
碩士 === 國立高雄應用科技大學 === 電子工程系碩士班 === 102 === A simple method of drawing relevant features of noise radiated from different targets is studied in this paper as is its application for ships recognition. Normally the radiation noise of a ship is composed of the broadband and narrowband components. The fo...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2012
|
Online Access: | http://ndltd.ncl.edu.tw/handle/5bk9p7 |
id |
ndltd-TW-102KUAS0393004 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-102KUAS03930042019-05-15T21:13:03Z http://ndltd.ncl.edu.tw/handle/5bk9p7 Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks 植基於水聲特性與類神經網路之船舶識別 Ching-Ching Ting 丁青青 碩士 國立高雄應用科技大學 電子工程系碩士班 102 A simple method of drawing relevant features of noise radiated from different targets is studied in this paper as is its application for ships recognition. Normally the radiation noise of a ship is composed of the broadband and narrowband components. The former is generated from the collision of a moving ship against its surrounding water body, characteristic of a resonant frequency, called critical frequency, the value of which is highly related to the speed of ship but usually falls in the range of 100 Hz – 1000 Hz. The magnitude of the broadband noise spectrum decays by 6 dB per octave attenuation below the critical frequency. On the other hand, the narrowband noise is mainly due to the operation of the propulsion system, which includes a variety of components such as the motor harmonic noise, the blades noise, and those induced by some mechanical instabilities. The narrowband noises highly related to the base frequency of the axial rotation. The Fourier Transform is utilized in analyzing ships noises, transforming the original time series signals into frequency spectrum and the top 10 frequencies in magnitude are picked out as the feature of the ship noises. These features are used as input to a back-propagation neural network system. A well-trained neural network based on the present study is regarded as capable of performing ships recognition rightly. The present method is practiced using the field data of radiation noises of 10 different fishing boats, performed with accuracy of more than 90%, showing a high practical value of the present study. 謝欽旭 2012 學位論文 ; thesis 60 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄應用科技大學 === 電子工程系碩士班 === 102 === A simple method of drawing relevant features of noise radiated from different targets is studied in this paper as is its application for ships recognition. Normally the radiation noise of a ship is composed of the broadband and narrowband components. The former is generated from the collision of a moving ship against its surrounding water body, characteristic of a resonant frequency, called critical frequency, the value of which is highly related to the speed of ship but usually falls in the range of 100 Hz – 1000 Hz. The magnitude of the broadband noise spectrum decays by 6 dB per octave attenuation below the critical frequency. On the other hand, the narrowband noise is mainly due to the operation of the propulsion system, which includes a variety of components such as the motor harmonic noise, the blades noise, and those induced by some mechanical instabilities. The narrowband noises highly related to the base frequency of the axial rotation. The Fourier Transform is utilized in analyzing ships noises, transforming the original time series signals into frequency spectrum and the top 10 frequencies in magnitude are picked out as the feature of the ship noises. These features are used as input to a back-propagation neural network system. A well-trained neural network based on the present study is regarded as capable of performing ships recognition rightly.
The present method is practiced using the field data of radiation noises of 10 different fishing boats, performed with accuracy of more than 90%, showing a high practical value of the present study.
|
author2 |
謝欽旭 |
author_facet |
謝欽旭 Ching-Ching Ting 丁青青 |
author |
Ching-Ching Ting 丁青青 |
spellingShingle |
Ching-Ching Ting 丁青青 Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
author_sort |
Ching-Ching Ting |
title |
Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
title_short |
Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
title_full |
Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
title_fullStr |
Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
title_full_unstemmed |
Vessel Identification based on Acoustic Characteristics and Artificial Neural Networks |
title_sort |
vessel identification based on acoustic characteristics and artificial neural networks |
publishDate |
2012 |
url |
http://ndltd.ncl.edu.tw/handle/5bk9p7 |
work_keys_str_mv |
AT chingchingting vesselidentificationbasedonacousticcharacteristicsandartificialneuralnetworks AT dīngqīngqīng vesselidentificationbasedonacousticcharacteristicsandartificialneuralnetworks AT chingchingting zhíjīyúshuǐshēngtèxìngyǔlèishénjīngwǎnglùzhīchuánbóshíbié AT dīngqīngqīng zhíjīyúshuǐshēngtèxìngyǔlèishénjīngwǎnglùzhīchuánbóshíbié |
_version_ |
1719110078297014272 |