Pest Prediction Revolution: Integrating Artificial Intelligence and Micro Weather Sensors for RealTime Data Driven Tropical Agriculture

Abstract

Tropical agriculture in emerging nations has a significant challenge from more unpredictable insect infestations brought on by climate change and environmental dynamics. In order to facilitate real-time agricultural decision-making, this project intends to create a pest attack prediction system based on the combination of artificial intelligence (AI) with microweather sensors. Installing microweather sensors in tropical agricultural areas, gathering environmental data and insect photos, and creating a hybrid CNN-LSTM model for analyzing and forecasting pest attacks are some of the techniques employed. Over the course of two growing seasons, the system was evaluated for forecast accuracy, agronomic effect, and economic analysis in a variety of tropical agroecosystems. In comparison to traditional approaches, the findings demonstrated that the AI-sensor system could lower the intensity of assaults by 58%, enhance the accuracy of pest attack prediction (F1-score 0.91; AUC-ROC 0.96), and reduce the consumption of pesticides by 67%. Furthermore, there was a notable rise in both economic efficiency and crop output. This study came to the conclusion that while large-scale deployment still necessitates infrastructure adaption and training, the combination of AI with real-time micro weather sensors has the potential to completely transform pest management systems in tropical agriculture.