AGROVATOR
The Future of Agriculture
The problem AGROVATOR solves
Maintaining Social Distancing has become a Must during after this Pandemic and might necessary to be followed even after. However, it might not be that easy for Agricultural Practices as it requires a huge human force to be involved during the cultivation of the crop. Hence we need to look for alternative methods of using Technology in Agricultural practices which make few steps of cultivation more efficient even with less human intervention. At the same time due to the transition in the economy, the government must come up with new methods for estimating the Minimum Support Price (MSP) as it is decided based on previous year statistics and it will not be efficient to follow after this economic transition. So we need to find suitable methods for deciding MSP.
Sowing the seed can however be done using modern equipment, we need to look forward to effective ways of disease diagnosis and also the estimation of crops which will be useful for both farmers and Government. So here is how our solution goes.
The surveillance area of the field is initially fed to the drone. The drone captures all the required data by moving over the field in a particular predefined path. Simultaneously this data will be sent to the server which will be analyzed using Machine Learning and Artificial Intelligence Algorithms to estimate various parameters. These parameters are used to know the health status of each crop. From those parameters, we can diagnose the disease of a crop more precisely (which can’t be detected by the naked eye) and take necessary measures to treat the plant. “Among all the crops, the yield of the cotton crop is reduced mainly due to diseases. Hence our project aims mainly at cotton crops".
The drone equipped with RGB, NIR cameras also can identify the crop. So, when it is surveilling the entire village area, it estimates the amount of land given to the various crops. This data helps the government in giving a better MSP (MINIMUM SUPPORT PRICE).
Challenges we ran into
Initially, it is really hard to make the Drone achieve a stable autonomous flight. Later using various additional stabilizing sensors onboard we managed to achieve a better stabilization. Using GPS we are able to achieve precise tracking of the Drone.
Next big problem is to gather the required data for feeding to the Machine Learning Algorithm we have developed. The necessity is to gather images of various plant species infected with different kinds of diseases. Moreover, the images must be captured in both the Visible and Infrared spectrum, which almost made us feel impossible. However, with got support for the Indian Institute of Rice Research(IIRR) and Prof. Jayashankar Agriculture University, we are able to gather almost all the required resources. Later with support from Mathworks we able to achieve our target of generating a regression pattern more efficiently.
Also, As the Cotton crop being the largest grown crop of south India(around 35% of whole crop produce) and also being one of the most disease-prone crops, we preferred to work on it. After studying a few samples, it is known that 50-55 plants out of 100 samples taken are infected with Cotton Leafhopper disease. Hence, at first, we confined our project to Cotton Leafhopper disease, which further can be extended to other diseases.