Accurate Prediction Model of Greenhouse Sensor Data based on Deep Learning Algorithm

Traditional greenhouse environmental monitoring systems mostly rely on simple algorithms for data processing. When faced with complex environmental factors, these methods often cannot accurately capture the temperature and humidity characteristics in the data, resulting in limited prediction accuracy. This article adopts a precise prediction model for greenhouse sensor data that combines Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) networks, aiming to improve prediction accuracy and the intelligence level of data analysis. The spatial features extracted by CNN are then fed into an LSTM network. The noise increased from 0% to 25%, and the accuracy of temperature prediction decreased from 90% to 72.82%, while the humidity dropped sharply from 88% to 58.62%. The research results indicate that the application of deep learning (DL) technology provides an efficient and innovative method for greenhouse environment monitoring, significantly enhancing the intelligence level and effectiveness of agricultural management.