With a high increase in the global population levels, the demand for resources mainly food that is highly necessary to sustain life is on the rise and is the biggest need of the future. The current outputs generated by the agricultural industries and lands is highly not sufficient to feed such huge numbers of people. To obtain this, we will have to increase the efficiency of the existing farming practices with the help of technology and this is exactly what we plan to do.
There are high suicidal rates among farmers in many places like India because of the crop destruction that happens due to certain plant diseases. This could be avoided and we could save the lives of many farmers as well as keep contributing to the global food supply if there was a well implemented method to keep check on each and every plant in the field. This is important as a single affected plant can still have the potential to spoil the whole field.
Ichigo which is a robotic rover that can go on almost all field terrains can help get the status of the entire field in terms of the plant health regarding the contraction of any plant diseases by the plant. The rover can be controlled from outside the field using a controller that also helps you click the picture of every desired leaf and later these images are then fed to a convolutional neural network to predict if the plant is affected by a disease or not. This would considerably revolutionise the lives of the farmers and could increase the efficiency of farming. The rover is also added with extra functionalities like giving the location of the diseased plant using GPS data and not destroying the crops using obstacle avoidance which is implemented through an ultrasonic sensor and is also equipped with a robotic arm that helps you cut off the weeds.
This project mainly has three major parts to it, the Sony playstation controller that sends the required signal, the computer which receives those signals and transmit them to the rover and also performs the image processing and the third part being the rover which follows the commands sent by the computer. The technical details for each of these parts are explained in detail in the next section.
Setting up the robot
- Connect the motors of the rover and the robotic arm to the motor driver and configure the pins accordingly on the arduino IDE. There is ample material on the web to refer regarding the connections of motors to motor drivers.
- Connect the motor drivers to the Sony board
- Connect the Sony camera module and the ultrasonic sensor to the Sony board and configure them on the arudino IDE
- Connect the Bluetooth module to the board for communicating via bluetooth
- Connect an appropriate power source to the motor drivers of the rover and the robotic rover
Setting up the Sony PlayStation Controller
Initially the PlayStation controller is paired to the computer via bluetooth and we make use of DS4 tools to establish the working of the controller with the windows operating system. The tool also helps us to map the keys and movements in the controller to specific keys on the keyboard. This is how we will be sending key presses to the robot.
Setting up the Computer
The computer is responsible for sending the signals to the bluetooth module in the robot using Processing. The code is given below in the code section. The computer also performs the image processing using Convolutional Neural Networks. We have made use of the nnabla libraries by Sony to perform the image processing. The implemented CNN is a Resnet model which is trained using a dataset (PlantVillage Dataset) with a collection of diseased and healthy leaf samples for specific types of plants.
The dataset used can be found here: https://github.com/spMohanty/PlantVillage-Dataset
The nnabla libraries are very well documented and the examples can be found in this link: https://github.com/sony/nnabla/blob/master/tutorial/by_examples.ipynb
The pictures are taken by the robot on the press of a particular button on the controller and this picture is stored in the sd card of the robot which after the whole journey throught the field can be uploaded to the computer which can later process the image and find out the diseased plants.