基于极限学习自编码器的水声信号目标识别方法

A Target Recognition Method for Underwater Acoustic Signal Based on Extreme Learning Autoencoder

摘要:

传统的机器学习方法在特征提取时容易受到主观经验的影响,导致对水声信号目标的识别准确率不高。而一般深度学习算法模型较复杂,通常具有训练耗时、计算复杂度高等缺点,极限学习自编码器具有很强的非线性处理能力,适合针对具有非线性特点的水声信号目标的识别,而且模型具有学习速度快,泛化能力强等显著优势,本文将极限学习自编码器算法应用于水声信号目标识别中,并与卷积神经网络、自编码器和极限学习机识别方法进行对比,结果表明,本文提出的方法对水声信号目标识别的准确率最优,且训练时间较短。

Abstract:

Traditional machine learning methods are easily influenced by subjective experience during feature extraction,Causing low recognition accuracy of underwater acoustic signal targets.Generally,deep learning algorithm models are complex,Training is time-consuming and computationally complex.Extreme Learning Autoencoder has strong non-linear processing ability,it is suitable for underwater acoustic signal target recognition with nonlinear characteristics.Moreover,the model has significant advantages such as fast learning speed and strong generalization ability.This paper applies the extreme learning autoencoder algorithm to underwater acoustic signal target recognition,and compare with convolutional neural networks,autoencoders,and extreme learning machine recognition methods.The result shows that the proposed method has the best accuracy in target recognition of underwater acoustic signals hand has a shorter training time.