Enhancing crop yield predictions with ensemble machine learning model using IoT environmental data and UAV vegetation indices : a case study of Maize | |
| Author | Nisit Pukrongta |
| Call Number | AIT Diss. no.ICT-24-02 |
| Subject(s) | Internet of things--Agricultural applications Agriculture--Data processing Crop improvement |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering in Information and Communication Technologies |
| Publisher | Asian Institute of Technology |
| Abstract | This dissertation introduces a weighted ensemble prediction model, PEnsemble4model, that combines several machine learning models to enhance the accuracy of maize yield at early growth stages. Utilizing both unmanned aerial vehicle (UAV) imagery and In ternet of Things (IoT) enabled sensors collected environmental data, as well as soil and plant nutrient attributes, the model offers a detailed, data-driven methodology for maize yield prediction. Given the increasing global demand for maize and the crop’s suscep tibility to weather fluctuations, the enhancement of predictive capabilities is crucial. The PEnsemble 4 model meets this demand by utilizing extensive datasets that include soil characteristics, nutrient levels, weather conditions, and UAV-captured vegetation imagery. It employs a combination of Huber and M estimates to analyze temporal varia tions in vegetation indices, particularly Chlorophyll-red edge index (CIre) and Normal ized difference red edge index (NDRE), which are key indicators of canopy density and plant height. Remarkably, the PEnsemble 4 model achieves an accuracy rate of 91%. It improves the timing of yield predictions from the traditional reproductive stage (R6) to the earlier blister stage (R2), thus facilitating more timely decision-making in agricul tural practices. Additionally, the model also capabilities extend to the water stress, crop stress , and disease detection, enhancing overall agricultural management. By integrat ing multi-modal data and machine learning technologies, the PEnsemble 4 model offers an innovative and effective approach to maize yield prediction. |
| Year | 2024 |
| Type | Dissertation |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information and Communication Technology (ICT) |
| Chairperson(s) | Attaphongse Taparugssanagorn |
| Examination Committee(s) | Vatcharaporn Esichaikul;Chaklam Silpasuwanchai |
| Scholarship Donor(s) | National Science and Technology Development Agency (NSTDA),Thailand |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2024 |