The Banana-Ai project is started!
Banana-Ai: Artificial Intelligence Based Modelling of Banana Plant Growth and Crop Quantity in Greenhouse and Creation of Dataset for Precision Agriculture Applications
The reason for choosing banana plant for the project is that the development of seedlings planted in the same area at the same time is not simultaneous. There is temporal and spatial variation. The average time between two flowering events in the same banana plant is also variable, unlike most crops whose phenology is synchronized with the seasons. These variations necessitate the construction of data-driven models for sustainable and efficient banana production.
It is known that plant growth, health and crop yield depend on soil moisture, temperature, and pH; light, air temperature and relative humidity in the greenhouse area should be monitored; and NPK (nitrogen, phosphorus, and potassium) values in the soil are important. Appropriate collection, recording and processing of these datasets is necessary for modelling, which is the objective of the project proposal.
Data sets related to the soil and the air in the greenhouse will be collected with the Internet of Things technology. LoRa based sensors will be adopted to measure the moisture, temperature, pH and NPK values of the soil, temperature and relative humidity of the air, and daylight reaching the greenhouse. Real-time data will be transferred from these sensors to the server using the LoRa protocol. The transmitted raw data sets will be processed to be used in artificial intelligence-based modelling.
Within the scope of the project, a dataset will be created with the data to be collected from the greenhouse during banana cultivation and the amount of product obtained. While creating this dataset, data will be recorded in natural greenhouse conditions where no intervention is made to the producer during the cultivation of bananas in the greenhouse. Within the scope of the project, no advice will be given to the producer, and no intervention will be made in the process. This dataset will then be used in studies such as monitoring the time-dependent change in the banana greenhouse, determining the need in the greenhouse, and estimating the amount of product to be obtained. In recent years, when precision agriculture practices have become widespread, data is of great importance in the success of studies in this field. The model is aimed to be a model that can predict with high coefficient of determination (>90%).