传感技术学报2024，Vol.37Issue(6)：1035-1040,6.DOI:10.3969/j.issn.1004-1699.2024.06.013
基于混合神经网络的多维视觉传感信号模式分类
Multidimensional Visual Sensing Signal Pattern Classification Based on Hybrid Neural Networks
摘要
Abstract
The poor classification accuracy of digital signals collected by sensors leads to the loss of key information.In order to improve the Reliability and effectiveness of sensing data,a multi-dimensional visual sensing signal pattern classification method based on hybrid neural networks is proposed.Combining convolutional neural network(CNN)and recurrent neural network(RNN),a hybrid neural net-work is constructed to represent the features in multidimensional visual data more effectively.Among them,convolutional neural network is responsible for denoising multi-dimensional spatial signals and extracting features,and recurrent neural network is responsible for fea-ture extraction of time-domain and frequency-domain signals.The hybrid neural network adjusts the weights of CNN and RNN by jointly training their respective parameters,and combines the features extracted from different levels to achieve multi-dimensional visual sensing signal pattern classification.The simulation results show that when using the proposed method for classification,the signal smoothness remains above 0.9,and the sensor signal classification results have a high fit with the actual results,effectively achieving multi-dimen-sional visual sensing signal pattern classification.关键词
传感器信号处理/信号模式分类/混合神经网络/视觉传感信号/卷积神经网络/循环神经网络/贝塞尔曲线Key words
sensor signal processing/signal mode classification/hybrid neural network/visual sensing signal/convolutional neural net-work/recurrent neural network/Bezier curve分类
信息技术与安全科学引用本文复制引用
陈威,蔡奕侨..基于混合神经网络的多维视觉传感信号模式分类[J].传感技术学报,2024,37(6):1035-1040,6.基金项目
福建省自然科学基金项目(2021J01318) （2021J01318）
