World Environment Congress (WEC)

35 / 2018-01-04 18:33:00

Reduction of pesticides usage in greenhouse – A Lean Neural Network Based Smart Robot

Smart Robot,ANN

Compared to spraying pesticides manually outdoors, the environment in greenhouse is more closed with higher temperature and humidity. Farmers indiscriminately use pesticides to avoid crop loss due to pest and fungal infections. Field study carried out in capsicum greenhouse farms show inefficient manual spraying due to the height of the plants ( about 9 feet ) leading to multiple applications with the hope pests are killed randomly.
An industrial robot with a 5 litre tank and six spray nozzles with variable droplet size mechanism was proposed and designed. The proposed system is controlled using a mobile phone app with input data from camera fixed in the robot to monitor plant and humidity sensor to identify the ideal droplet size that has to be sprayed. A software to analyze the data using a lean neural network to obtain the spraying parameters is proposed. Simulation results show pesticide savings by 20%. Real time experiments are being carried out and data is being measured. The effectiveness of this platform is shown by the systems ability to successfully navigate itself down rows of a greenhouse, while the pesticide spraying system efficiently covers the plants evenly with spray in the set dosages computed. The exposure of labors to toxic pesticides are also reduced.