From Sensors to Artificial Intelligence: Field Excitement in the Banana-AI Project
Engineering students at Alanya Alaaddin Keykubat University (ALKU) are redefining traditional farming methods with smart technologies through the Banana-AI project, carried out with the support of TÜBİTAK 1001 (Project No: 125E197). In this innovative study, conducted under the supervision of Assoc. Prof. Dr. Akın Oktav, a team of nine students — five from Computer Engineering and four from Electrical-Electronics Engineering — is transforming banana greenhouses into digital data factories.
The future success of the project is being built on the solid foundations we have laid in the laboratory. In the comprehensive validation tests we conducted before deploying the sensors to the greenhouse, we meticulously examined, one by one, the measurement behavior of 72 sensors that measure critical data such as soil moisture, pH, NPK, CO2, and light. We optimized our system by filtering out devices that fell outside the expected accuracy range, and we prepared industrial panels and uninterruptible power supplies to ensure the seamless transfer of data to the ThingsBoard platform. With this reliable infrastructure experience gained during the laboratory phase, we are now preparing to move our entire system to the field — that is, to the greenhouses — to implement smart agriculture applications under real production conditions and to achieve our targeted productivity results.
The most striking aspect of the project is the multi-layered imaging techniques used in analyzing plant health. While RGB cameras track the physical development of the plants, multispectral cameras and drones capture details invisible to the human eye. The analysis results obtained from these acquisitions — optimized with Teflon calibration plates and custom-built setups — are combined with the data coming from the soil sensors on the ThingsBoard platform to pinpoint the plants' instantaneous needs.
This work prevents unnecessary irrigation by ensuring maximum water savings and aims to predict crop yield with an accuracy of 90% or higher using artificial intelligence models. The massive dataset to be obtained at the end of the project will constitute a unique resource for agricultural artificial intelligence research in Türkiye. Making a strong contribution to the digitalization vision of regional and national agriculture, the Banana-AI project will continue to pioneer the farming practices of the future for 36 months.
