loading page

Underwater Acoustic Target Recognition Using Spectrogram ROI Approximation And Mobilenet One-Dimensional And Two-Dimensional Networks
  • Hassan Akbarian ,
  • Mohammad Hossein Sedaaghi
Hassan Akbarian
Sahand University of Technology, Sahand University of Technology

Corresponding Author:[email protected]

Author Profile
Mohammad Hossein Sedaaghi
Author Profile

Abstract

In this paper, we proposed a new model for underwater acoustic target recognition, which is based on the deep learning methods (CNN, VGG19, and MobileNet) with modified mechanisms at the end of the network in the ShipsEar dataset. we show that the automatic target recognition of ship-radiated noise with deep learning methods has a relatively more appropriate and reliable performance. This is significant because This research has intensively tried to achieve a proper balance between the accuracy and speed of the networks, so that with the relative improvement of the classification accuracy,they have noticeably reduced the number of parameters and the amount of calculations. The proposed models presented in this paper will run on portable devices and mobile phones. The proposed method aided in the Sonar system’s acoustic target classification and recognition.