AI driven farmbot for crop health monitoring and disease detection

Abstract

The increasing population in urban areas is associated with a number of problems like limited land for carrying out agriculture which leads to over dependence on buying food from the market. With an increase in food prices, the cost of living in urban areas increases leading to hunger. It can be seen that the land use pattern in urban centres is greatly affected leaving land used for agriculture in urban centres with a risk of being used for other activities like settlement and industrialization. A high growth rate of urbanization reduces the agricultural production resulting in the loss of potential crop yield due to the transformation of productive land to its non-productive usage. This reduces soil quality in terms of soil nutrients and aeration which leads to food insecurity. The farmbot project was developed to make farming possible in urban areas where people have limited land, knowledge about farming and have busy schedules thus limited time to practice it effectively using their mobile devices. In this report we have improved the existing farmbot by introducing three systems that is; disease detection system, the water stress determination and crop growth monitoring systems which are all equally important for the crop health and high productivity. We aim at using locally available materials so that we can make the farmbot affordable to people who would like to have a small garden at their homes for constant food supply. With no improvement in the existing farmbot, there will be continuous poor crop yield from small-scale farms in urban areas hence low food production and low food supply to meet the food demand. The cost of living of people in urban areas will increase due to over dependency and reliance on buying food from market places at high prices. The low-income earners will not be able to afford the heightened food prices leading to malnutrition and hunger. Design of components was made basing on the load requirements, the material used and the purpose. Components were fabricated, others 3D printed and later they were connected and the system was implemented following the circuit diagrams generated from circuito.io and proteus and flow charts to come up with algorithm of the system and programmed in raspberry pi3 and Arduino for the hardware section and various languages for the software section were used such as C++, MATLAB, python and JavaScript. The system was tested at different levels, for example unit testing, integral testing and finally system testing and results discussed accordingly, indicating the efficiency of the system. An economic analysis was carried-out to establish the viability of the project, the Net Present Value was greater than 1 showing that the project is viable, the payback period was calculated giving a period of about one and a half years. The conclusions were drawn, which indicated the efficiency of the system in improving the crop yield by detecting disease, determining water stress and monitoring growth of the crop with minimal human interaction. Prototype was developed using available materials and components and it was able to move from one crop to the next crop and as well able to pick the desired tool and use it at the desired points. The quality of the crop produced with an improved AI driven farmbot showed great improvements than the one produced with the already existing farmbot in terms of the leaf appearance.