Estimation of Plant Growth Environment Temperature based on Plant Electrical Signals

Electrical signals are an important type of signals in plants, which can not only regulate normal physiological activities, but also reflect the current environmental temperature of plants to a certain extent. The plant electrical signal was collected and the Peak value, Peak to peak and Mean value were extracted as eigenvalues, and the BP neural network was trained to obtain the correspondence model between the eigenvalues of the signal and the ambient temperature. The mean square deviation (MSE) and correlation coefficient (R) of the trained BP neural network model were 0.0058 and 0.9408, respectively. In the re-collected data, 90% of the estimated values fall between 0~1, and the judgment effect is good, which proves that the model has a high accuracy of judgment. In the future, plant electrical signals can be combined with neural network models to detect temperature, humidity and other parameters in plant growth states, which has high application value in smart agriculture.