Temperature Real-time Monitoring System Based On Temperature Sensor Array And Deep Learning

Body temperature is an important measure of the condition of the body.Its accuracy directly affects the diagnosis,treatment and care of the disease.At present,the measurement of body temperature in hospitals mainly uses mercury thermometers or hand-held electronic thermometers,but it cannot meet the demand of modern medical care for high-precision,continuous real-time monitoring of inpatients.For this reason,this article designed a real-time temperature monitoring system based on temperature sensor array and deep learning.In order to realize the wireless real-time monitoring of the patient's body temperature,this subject mainly studies the problems of data acquisition accuracy and data acquisition reliability in the wireless monitoring system.In order to improve the accuracy of body temperature data acquisition,according to the error characteristics of DS18B20 commercial temperature sensor in the human body temperature range,the temperature sensor array and data fitting algorithm are used to achieve high-precision body temperature measurement,and the algorithm is used in the coordinator to verify the improvement of the algorithm.The actual effect of precision.In order to ensure the accuracy of wearable body temperature measurement and effectively identify external disturbances in the acquisition process,this paper proposes and compares Kalman filter algorithm,improved Kalman filter algorithm based on support vector machine,and artificial neural network algorithm filtering effect.The application of the support vector machine classification algorithm to anti-interference filtering of body temperature sensors.The algorithm cuts a large amount of continuous historical body temperature data into samples with four eigenvalues.Technicians and medical staff can label labels according to the characteristics of the body temperature measurement system's interference signals,and divide the samples into training and test sets according to a certain proportion.Support vector The machine classification algorithm uses the training set for model training and validates the model classification accuracy based on the test set.Adjusting the model training parameters,the simulation verified that the algorithm's recognition accuracy of the interference signal in the body temperature monitoring system meets the system design requirements.Based on the hardware platform and PC client software designed in this project,a high-precision continuous real-time measurement of multi-node body temperature and real-time display and storage of body temperature data are realized.By automatically performing the Reset operation on the current network state of the node,automatic switching between routing sub-nets is implemented to ensure continuous data transmission.The test results show that the measurement accuracy of this system can reach ±0.1°C in the temperature range of 35°C-39°C;the recognition accuracy of the interference signal is 98.75%.