Multivariate Multi-Step Agrometeorological Forecast Model for Rapid Spray

The timing of spray application plays an essential role in daily pesticides management. Proper wind speed, air temperature, and relative humidity are the main external factors to improve the efficacy of pesticides, reduce the amount of spray drift and environmental pollution. Very few previous studies have focused on the need for short-term weather prediction in rapid spraying decisions. In this paper, a Convolutional-LSTM encoder-decoder (ConvLSTM-AE) hybrid model for multivariate output and multi-step prediction with short time intervals is proposed to predict these three agrometeorological variables in advance. This model can predict daily weather conditions at 15-minute interval and track the changes of time-varying systems effectively. This method was also compared with other methods such as CNN, multi-head CNN, LSTM encoder-decoder, and CNN-LSTM encoder-decoder models. The results show that the proposed model outperforms other models and is suitable for daily weather forecasts in a short time interval. The obtained rapid and accurate prediction results provide a reliable basis for precise spray timing in actual farming.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.