没有合适的资源?快使用搜索试试~ 我知道了~
资源推荐
资源详情
资源评论
中北大学学位论文
Design of ECG Intelligent Monitoring System Based on Neural
Network
Abstract
With the accelerated pace of people's lives and increased pressure, coupled with the
growing trend of population aging, the incidence of heart disease and other diseases is
increasing, seriously threatening people's lives and health safety, and people are gradually
paying attention to their own health, and the demand for medical devices for active monitoring
is increasing. Most of the existing ECG monitoring systems are limited to the transmission,
reception and display of ECG signals, and without the help of medical professionals, the
subjects cannot understand their own heart health status. In this paper, we combine deep
learning with ECG monitoring system and design a neural network-based ECG intelligent
monitoring system. In addition to the basic acquisition, transmission and display, we add the
intelligent diagnosis function of ECG data, which can make preliminary diagnosis of ECG data
and facilitate the subjects to know their heart health status in time. The main contents of this
paper are as follows:
Firstly, the paper analyzes the characteristics of ECG signals as well as the application
scenarios and needs of the intelligent monitoring system, and introduces the general scheme of
the ECG intelligent monitoring system. Then, the design of the ECG acquisition device and the
ZigBee transmission module are introduced. The processing circuit of the ECG signal uses the
ADS1292R chip, the master control module uses Arduino, and the ZigBee module uses the
CC2530 chip.
Then, the noise and pre-processing methods of the ECG signal are introduced, and the
Finite Impulse Response (FIR) filter used is described. After that, the paper describes in detail
the algorithm to perform ECG intelligent recognition, and the CNN model is finally selected
by designing, testing and studying the models of Long-Short Term Memory (LSTM) network,
中北大学学位论文
Bi-directional Long Short-Term Memory (BiLSTM) network, Convolutional Neural Network
(CNN) and the combination of CNN and BiLSTM, and optimizing the activation function by
testing and studying different model. The final recognition rate of ten ECG signal types,
including normal heartbeat and left bundle branch conduction block, reached 99.50% on the
test set. Next, the ECG intelligent monitoring system processing platform is introduced, and it
is developed using PyQt based on Python to make it more compatible and portable.
Finally, the system functions were tested, and the test results showed that the system can
realize the pre-processing of data and preliminary diagnosis in addition to the basic functions
of ECG data acquisition and transmission, and the real-time performance of the model
algorithm is ideal, which meets the design requirements.
Keywords: ECG, Convolutional Neural Network, Intelligent Monitoring, ZigBee
中北大学学位论文
I
目 录
1 绪论
1.1 研究背景及意义 .................................................. 1
1.2 国内外研究现状 .................................................. 3
1.2.1 可穿戴健康监测系统 ......................................... 3
1.2.2 心电信号智能识别 ........................................... 6
1.3 研究内容与结构安排............................................... 7
2 心电智能监测系统总体方案
2.1 心电信号特征分析 ................................................ 9
2.2 心电智能监测系统需求分析 ........................................ 11
2.3 系统总体设计 ................................................... 11
2.4 无线通信方案选择 ............................................... 13
2.5 本章小结 ....................................................... 14
3 心电智能监测系统采集装置设计
3.1 采集装置基本组成 ............................................... 15
3.2 信号采集处理电路 ............................................... 15
3.3 主控模块设计 ................................................... 18
3.4 基于 ZigBee 的数据传输模块设计 ................................... 19
3.4.1 ZigBee 协议结构 ............................................ 19
3.4.2 ZigBee 网络拓扑结构 ........................................ 20
3.4.3 ZigBee 协议栈 ............................................. 21
3.4.4 ZigBee 软件设计 ............................................ 22
3.4.5 ZigBee 硬件设计 ............................................ 23
3.5 本章小结 ....................................................... 24
4 心电信号处理与识别算法研究
4.1 心电数据集介绍 ................................................. 25
中北大学学位论文
II
4.2 数据预处理 ..................................................... 25
4.2.1 心电信号中的常见噪声 ...................................... 25
4.2.2 除噪方法 .................................................. 26
4.3 R 峰定位方法及测试 .............................................. 28
4.4 识别算法设计 ................................................... 29
4.4.1 神经网络简介 .............................................. 29
4.4.2 长短期记忆网络简介 ........................................ 30
4.4.3 卷积神经网络简介 .......................................... 32
4.4.4 长短期记忆网络模型设计 .................................... 34
4.4.5 卷积神经网络模型设计 ...................................... 35
4.4.6 CNN-BiLSTM 模型设计 ........................................ 37
4.4.7 训练及测试结果 ............................................ 37
4.4.8 激活函数对模型的影响 ...................................... 40
4.5 本章小结 ....................................................... 43
5 心电智能监测系统处理平台设计与试验测试
5.1 心电智能监测系统处理平台设计 .................................... 44
5.1.1 处理平台总体设计 .......................................... 44
5.1.2 处理平台 UI 界面设计 ....................................... 45
5.1.3 处理平台逻辑设计 .......................................... 47
5.1.4 心率计算 .................................................. 49
5.2 试验测试 ....................................................... 49
5.2.1 系统搭建 .................................................. 49
5.2.2 试验数据分析 .............................................. 51
5.2.3 异常心电数据测试 .......................................... 52
5.2.4 神经网络模型测试 .......................................... 53
5.3 本章小结 ....................................................... 54
6 总结与展望
剩余63页未读,继续阅读
资源评论
2201_75761617
- 粉丝: 18
- 资源: 7339
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功