A neural network based seafloor classification using acoustic backscatter

Abstract

This paper presents a study results of the Artificial Neural Network (ANN) architectures [Self-Organizing Map (SOM) and Multi-Layer Perceptron (MLP)] using single beam echosounding data. The single beam echosounder, operable at 12 kHz, has been used for backscatter data acquisitions from three distinctly different seafloor's from the Arabian Sea. With some preprocessing of the snapshots, the performance of the SOM network is observed to be quite good. For unsupervised SOM network, only single snapshot is used for the training, and number of snapshots for subsequent testing of the network. Feature selection from ASCII data is an important component for a supervised MLP based network. Four selected features are used for training the network. The test results of the MLP based network are also discussed in the text.