By Alistina Shrestha
May 24, 2024: Our groundbreaking research titled, “Enhancing Crop Yield Predictions with PEnsemble 4: IoT and ML-Driven for Precision Agriculture” has been published in Applied Sciences as part of the Special Issue on AI, IoT, and Remote Sensing in Precision Agriculture. Nisit Pukrongta, a Ph.D. student in the Information and Communications Technologies program, under the guidance of his advisor, Dr. Attaphongse Taparugssanagorn, has successfully introduced a new model, PEnsemble 4, as the cornerstone of this study. PEnsemble 4 uniquely integrates multiple machine learning models, unmanned aerial vehicle (UAV) imagery, and Internet of Things (IoT) based environmental data. This model aims to significantly improve the accuracy of predicting maize yield, a crucial factor in agricultural practices.
Figure 1: Nisit Pukrongta (Right) with his advisor, Dr. Attaphongse Taparugssanagorn
The PEnsemble 4 model improves prediction accuracy by analyzing soil quality, nutrient composition, weather conditions, and UAV-captured vegetation images. The model uses Huber and M estimates to analyze patterns in vegetation indicators, especially CIre and NDRE, which serve as reliable indicators of canopy density and plant height. It demonstrates a remarkable accuracy rate of 91%.
Figure 2: Aerial view of the maize field
This research addresses a critical gap in crop health prediction by presenting an innovative approach that utilizes the combined power of the IoT and AI. The groundbreaking system proposed in this study stands as a pioneering solution that has the potential to revolutionize the agricultural landscape. It aims to achieve the following objectives:
- Continuous crop monitoring: Develop a comprehensive and continuous monitoring system that provides real-time insights into various environmental and crop-specific parameters. Integrating IoT devices into agricultural practices enables the collection of dynamic data, helping farmers and agronomists with an unobstructed view of their crops’ health and growth.
- Precise predictive modeling: Use advanced AI models to analyze data from IoT devices carefully. These models are designed to give timely and accurate information about crop health and growth patterns. By using the power of predictive modeling, farmers can take specific actions to use resources better and improve crop quality.
Figure 3: Nisit Pukrongta in the field
The key insights of the research are:
- The study introduces a new method for farmers to improve crop output, management, and protection using advanced technology. It enables the farmers to predict maize grain production as early as day 79 of the R2 stage, much earlier than traditional methods. The early prediction capability of the PEnsemble 4 model can significantly enhance farmers’ decision-making and resource management.
- It utilizes IoT and machine learning technologies to integrate various data sources, such as soil quality, weather conditions, and UAV-captured imagery, to enhance the accuracy of yield forecasts.
- The PEnsemble 4 model enables early detection of crop stress and disease outbreaks, helping farmers take quick action to protect their crops.
- The model helps farmers maximize productivity while minimizing environmental impact and also plays a crucial role in ensuring global food security. By contributing to sustainable farming practices, this model has the potential to significantly enhance our ability to meet the world’s growing food demands.
Figure 4: Nisit Pukrongta collecting the data from the site
Figure 5: Data Collection via UAV imagery and IoT
The research has achieved several Sustainable Development Goals: Goal 2: Zero Hunger – boosting crop yields to ensure food security by providing early and accurate yield predictions; Goal 9: Industry, Innovation, and Infrastructure – harnessing technology for sustainable agriculture by integrating IoT and machine learning into crop monitoring; and Goal 13: Climate Action – mitigating risks by optimizing resource usage and reducing environmental impact through precise predictive modeling. These contributions demonstrate the potential of the PEnsemble 4 model in addressing global challenges and promoting sustainable farming practices.
In summary, the outcomes of this study are of utmost importance for agricultural practices, particularly in precision agriculture. Identifying specific growth stages and vegetation indicators correlating with seed weight offers invaluable guidance for farmers. The farmers can optimize seed yield by concentrating their efforts on irrigation and nutrient management during these critical periods. Moreover, the study highlights the importance of understanding the correlations between environmental factors, including weather and soil conditions, and seed weight. This knowledge helps farmers make informed decisions regarding irrigation, pest control, and other environmental management practices, enhancing crop production.
In conclusion, this research offers a promising future for precision agriculture, where technology and data-driven insights can revolutionize agricultural practices. We congratulate Nisit Pukrongta on this remarkable achievement and wish him continued success in his future endeavors.
Read the complete research in the link below:
Enhancing Crop Yield Predictions with PEnsemble 4: IoT and ML-Driven for Precision Agriculture
