Monitoring the Physiological Conditions of the Azolla Growth using Artificial Technology with Raspberry Pi

Abstract

The optimal conditions for cultivating Azolla in Malaysia are continuously characterised by a consistently humid climate, with an average daily temperature ranging from 21°C to 32°C. The optimal relative humidity for Azolla growth falls within the 85-90% range. However, excessively high humidity combined with elevated temperatures promotes the proliferation of insects and fungi. A monitoring system was required to help harvest the best possible quality of Azolla. This study proposes that artificial intelligence (AI) be used to monitor the physiological conditions of its growth. Two crop containers were differentiated with added Phosphorus fertiliser and without fertiliser. For eight days, both containers were monitored. The system used a Raspberry Pi camera to capture images of Azolla plants and pests, and a machine-learning model was trained on a dataset of healthy and unhealthy Azolla plants and pests. Raspberry Pi 4 Model B has been used and has made quick progress. TensorFlow and OpenCV are used to extract features from the images. Results show that using TensorFlow Lite and OpenCV in this project has improved crop yield and identify diseases and pests early. The developed AI system enables the assessment of Azolla's condition and the identification of the pests responsible for damaging it. This project helped to contribute to future research on enhanced techniques and technologies for monitoring crops more efficiently.