Weed detection in crops using aerial vehicle and machine learning technique

Agriculture is the largest sector, covering 47.03% of land area of Pakistan and almost 45% of workforce engaged within this sector. Improper management, use of old techniques and technology leads to decrease in crop yield. Weeds growth among crops has great impact on yield. Proper monitoring and man

Project Title

Weed detection in crops using aerial vehicle and machine learning technique

Project Area of Specialization

Electrical/Electronic Engineering

Project Summary

Agriculture is the largest sector, covering 47.03% of land area of Pakistan and almost 45% of workforce engaged within this sector. Improper management, use of old techniques and technology leads to decrease in crop yield. Weeds growth among crops has great impact on yield. Proper monitoring and management of crops will help in improving crop productivity and efficient use of resources. The main objective of our project is to build an autonomous hexa-copter that will be used for detection and removal of weeds among crops using image processing and machine learning technique, which will help in improving crops health. The execution of algorithms is in Python language for efficient and faster processing.  The hardware integration of various components are as shown in Fig. 1.

                                                           Fig. 1: Layout of project

Project Objectives

• Real-time weeds detection in a crop using Machine Learning.
• Hexa-copter with Camera Interface and Laser Weed cutter.
• Autonomous operation
• Weed removal.

Project Implementation Method

Hardware: A hex-copter drone is being designed and tested for flight over a crop field area. After the deployment next step is to set the GPS coordinates of field in code and its execution in RPI terminal. The hexa-copter will start its autonomous operation, flying above crops. On detection of weed, the flight will stop for some time and perform removal operation.

Software: The programming section contains algorithm for weed classification and detection using Python language. Softwares used in project are as follows:

• Raspbian RPI terminal
• Arduino Integrated development Environment (IDE)
• Anaconda Navigator and Python Programming

The machine learning section contains weeds dataset containing images of various weeds under consideration. Different libraries are used such as OpenCV, Drone-kit for implementation of project. In Removal section, a suitable laser is used. The general implementation of project can be visualized through the block diagram as shown in Fig. 2.

Fig. 2   Hardware and software interface

Proposed Algorithm

Fig. 3: Flowchart of Proposed Algorithm

Fig 4: Weed Removal Process

Benefits of the Project

• Time efficient.
• Autonomous operation.
• Weed detection and removal.
• Increase in yield.
• Easy deployment.

Technical Details of Final Deliverable

Laptop and Raspberry pi are connected through Wi-Fi channel without need of Internet. Now the RPI terminal with compiler shows up in laptop, which enables control of Raspberry. Raspberry and Arduino are connected to each other through serial communication. Arduino sends PWM signals to Hexa-copter received by APM flight controller. Outputs of APM controls ESCs which is connected to Motors. On receiving signals from APM the ESC turns the motor ON and Hexa-copter starts flying. Last step is the image processing which starts when drone starts flying autonomously above the crops. After recognizing its target, the Hexa-copter will stop for moment and turn the laser on to burn out weed.

Final Deliverable of the Project

Hardware System

Core Industry

Agriculture

Other Industries

Core Technology

Artificial Intelligence(AI)

Other Technologies

Others

Sustainable Development Goals

Decent Work and Economic Growth, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com