Autonomous Rover for Plant Disease Identifications

INTRODUCTION

Sri Lanka is a fertile tropical land located in the Indian Ocean with the potential for the cultivation and processing of a variety of crops. Hence, agriculture has become one of the main component of the Sri Lankan Economy. Sri Lanka is popular for cultivation of Rice, Tea, Rubber and large number of other important crops. Hence, the motivation of this project is to develop an Autonomous Rover which could navigate through these plantations and detect plant diseases using image processing and artificial intelligence techniques and notify the farmers with the required information.

PROJECT DESCRIPTION

This project will be carried out by designing and implementing a DonkeyCar robotic platform which will initially be programmed to navigate through an entire cultivation automatically. The path of navigation would initially will be programmed. DonkeyCar platform would be ideal for this situation, hence it is equipped with excellent safety and navigation features to move through plantation lands. For a much larger plantations GPS Module will be used to set the way points of navigation.

When the rover is navigating towards the plantation, it will continuously monitor the leaves of the plants on either sides using two Camera modules. The images taken through these cameras will be pre-processed using certain image processing techniques and the resulting images would be checked for diseases using a neural network.

Then finally the obtained data would be transmitted in to an IoT platform where the farmers can access through their mobile phones and observe the type of the diseases and the exact locations of the diseased plant and possibly the remedial actions that has to be taken.

PROJECT BREAKDOWN

STAGE 01

Stage 01 of the project will be focused on developing the hardware components. DonkeyCar robotic platform will be optimized based on the objectives and required components will be interfaced in to the platform. Chassis would be developed with obstacle avoiding sensors and two cameras at either sides to observe two sides different sides of the farm at once while navigating through it. GPS module will be interfaced to the chassis to set the path for the robot to navigate once operational.

STAGE 02

In Stage 02 of the project, multi spectral images of plant leaves will be taken continuously while navigating through the filed. These images will be pre-processed to reduce salt and pepper noises and certain image processing techniques will be added upon the requirement.

STAGE 03

In 03rd Stage of the project a Convolution Neural Network will be implemented using deep learning algorithms to recognize different diseases in plant leaves. In here classifiers will be developed for different diseases with large number of sample images and their respective diseases.

STAGE 04

Finally, the project will be IoT enabled through a web interface, which could be accessed by the farmers which will display the important parameters of the cultivation and the detected diseases, location of the diseased plants (which will be superimposed on a map of the plantation) and also possible remedies and precautions against such diseases.

SCALABILITY

The most important aspect of this project is that it can be continuously improved to include new diseases and train the system continuously with new sample images. System can be further develop to autonomously pluck the disease infected leaves using certain actuators to prevent the spread of the disease as well.

POWER EFFICIENCY

Power Efficiency is another major aspect of this project, this robot will be powered using Li-Ion or Li-Po Batteries which will be charged by a solar panel which will be interfaced to the robot chassis as well. Hence, the batteries can be charged while the robot is operational and it will improve the efficiency of robot tremendously.

CONCLUSION

Final objective of this project is to improve the existing agricultural system in Sri Lanka and through out the world by implementing advanced technological developments in order to make the existing processes much more efficient.