Marine mammal sounding real-time identification method based on convolutional neural network

The invention relates to a marine mammal sounding real-time identification method based on a convolutional neural network, which belongs to the field of signal processing. The method comprises the following steps of: performing standardization processing on a segment of signals, generating a time-frequency color image of the signal through fast Fourier transform, then labeling animal sound signals in the time-frequency image, and finally, sending the data into the convolutional neural network for training to obtain an identification model. According to the method, the marine mammal sound signals in the underwater sound data are recognized in one step by adopting the identification model constructed by the convolutional neural network. The method provided by the invention has high recall rate and accuracy and high processing speed, and can meet the requirement of real-time identification. 本发明涉及一种基于一种卷积神经网络的海洋哺乳动物发声实时识别方法，属于信号处理领域，首先将一段信号进行标准化处理后通过快速傅里叶变化产生信号的时频彩色图像，随后对时频图像中的动物声信号进行标注。最后将数据送入卷积神经网络进行训练以得到一个识别模型