Smart farming using IOT and machine learning in Jimma city Jiren kebele

Abstract

Agriculture is one of the issues that no-way runs out to be examined. Since farming is one of the main sources of livelihood of the pastoral population. Thus, the construction of agrarian data grounded internet of effects is veritably important to do. In Ethiopia collecting information like fertility of the soil, rainfall, growth of crops, temperature, downfall and information regarding colony of seeds, etc. can be collected with the help of IOT. It helps the growers to gain information regarding all farming conditions. With the help of internet technology, farming processes can be covered through detectors, smart cameras, mobile operations and bias like mini chips. Through IOT the automated internet technology, farmers can utilize farming resource efficiently. In Jimma zone farmers rely on their experience to sow seeds on the soil. Without knowing what types of crops can be grown in the soil accurately, they cultivate the land and grow crops based on usual way of farming. Farmers do not try to crop new species in their land because if it fails to grow, they will lose revenue from farming. So, if there is a means that will provide the information to let them grow potential crops that will yield optimum production will be helpful to them. In addition the soil they cultivate may have enough minerals to grow surplus crops that are not grown in the region. Having the mineral content information of the soil in advance enhances optimum crop production. Previously many researches have been done in crop prediction in another country. But up to my knowledge there is no existing crop prediction research that is done on Jimma zone in Jiren kebele. The data that is collected and the model that is built aimed for this kebele only. This research project is prediction of the types of crops that will grow in Jimma zone Jiren kebele. Based on the soil mineral and environmental condition data of the kebele a machine learning model was built to predict the types of crops that can be grown in the area. Different sensors such as moisture sensor, temperature sensor, pH sensor and NPK (nutrient sensor) were

employed in the soil to get mineral and environmental condition data. And then decision tree algorithm was used to build a machine learning model that will predict the types of crops that will be grown in the soil. The average accuracy of the model in decision tree after 22 times run is 0.97.