• Abstract
• What are we doing
• Technology Stack
  • Programming Language
  • Editor
  • Libraries
• Dataset
• Further Reading
• Report
• Outputs
• Team

The agricultural industry continually seeks innovative solutions to enhance productivity, address labor shortages, and optimize crop harvesting processes. In this context, the development of an autonomous tomato harvesting robot equipped with machine learning capabilities presents a promising approach. Our project aims to design and implement an autonomous tomato harvesting robot with a 6-degrees-of-freedom (6DOF) robotic arm, enabling precise and gentle fruit detachment to minimize damage and waste. At its core lies a sophisticated deep learning neural network, a lightweight model tailored to meet the real time application requirements of CNNs in low-cost embedded systems such as IoT applications. Employing the state-of-the-art YOLOv4 tiny object detection with the Darknet framework, the system accurately identifies and locates unripe, semi-ripe, and ripe tomatoes in the field, surpassing traditional methods and ensuring efficient and accurate harvesting. By combining the dexterous 6DOF arm with vision from an Insta360 camera and YOLOv4's tiny yet robust detection capabilities in real-time, along with integrating a TOF sensor for depth calculation, our project paves the way for a future of autonomous and intelligent tomato harvesting. This innovative approach holds immense potential to boost agricultural efficiency and contribute to a more sustainable food production system.

• To reduce the labour shortage on tomato harvesting.

• Programming Language

  • Python

• Editor

  • Visual Studio Code

• Libraries

  • Darknet: The framework used for implementing YOLOv4 for tomato detection.
  • OpenCV: Used for image processing and computer vision tasks.
  • Matplotlib: Used for data visualization.
  • Pandas: Used for data manipulation and analysis.
  • NumPy: Used for numerical computing.

• Dataset

  The project involves the use of datasets related to tomato harvesting and machine learning.

• kaggle-Tomato Detection

• LaboroTomato

• TomatOD

For additional reading, we recommend the following resources:

• Python Machine Learning
• Introduction to 6DOF Robotic Arms with Python
• YOLOv4tiny

Detailed project report

The project team includes:

• Mausam Gurung

  • GitHub: MausamGurung
  • LinkedIn: Mausam Gurung

• Shubham Bista

  • GitHub: ShubhamBista
  • LinkedIn: Shubham Bista

• Symon Shrestha

  • GitHub: SymonShrestha
  • LinkedIn: Symon Shrestha

• Shalon Maharjan

  • GitHub: ShalonMaharjan
  • LinkedIn: Shalon Maharjan