Classification of Underwater Objects Based on Probabilistic Neural Network

Classification of underwater objects remains challenging and significant problem because of the complexity of underwater environments. In this paper,a probabilistic neural network (PNN) is used as a classifier to the automatic classification of underwater objects. Firstly, a process of multi-field f...

Full description

Saved in:
Bibliographic Details
Main Authors: Jie Tian, Shanhua Xue, Haining Huang
Format: Conference Proceeding
Language:English
Subjects:
Character recognition
Feature extraction
Lakes
Neural networks
Noise robustness
Probabilistic Neural Network
Support vector machine classification
Support vector machines
Target recognition
Time domain analysis
Time frequency analysis
Underwater object classification
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Classification of underwater objects remains challenging and significant problem because of the complexity of underwater environments. In this paper,a probabilistic neural network (PNN) is used as a classifier to the automatic classification of underwater objects. Firstly, a process of multi-field feature extraction is employed to construct a feature vector.The multi-field feature extraction involves time-domain analysis, time-frequency distribution, spectra and bispectra analysis. Underwater target classification can be considered as a problem of small sample recognition, because samples acquired under different conditions often exhibit different clustering characteristics. Probabilistic neural network is chosen to discriminate underwater objects because of its simplicity, robustness to noise, and nonlinear decision boundaries. The PNN classifier is contrasted with a Gaussian classifier and a support vector machine (SVM) using lake or sea trial data. Experimental results indicated the PNN classifier is appropriate to this problem.
ISSN:2157-9555
DOI:10.1109/ICNC.2009.362