ABSTRACT
Agriculture today is rapidly embracing automation and IoT technologies to boost productivity and make farming smarter. This project introduces an Automated Vegetable Picker Machine that uses IoT and AI to detect and harvest ripe vegetables with better efficiency. The system combines an ESP32-CAM module for image capture, a color sensor for ripeness detection and a 4-DOF robotic arm to carefully pick the vegetables without causing damage. A lightweight AI model, trained using platforms like Google Teachable Machine or TensorFlow, is deployed on the ESP32-CAM to classify vegetables.
Ripeness in real time. Once a ripe vegetable is identified, the Node MCU processes the data and sends commands to the robotic arm to carry out the picking action with precision. The entire setup is connected to a mobile application via IoT, allowing farmers to receive live updates and notifications on harvesting progress. This level of automation reduces dependence on manual labor, speeds up the harvesting process and minimizes post-harvest losses caused by inconsistent human handling. The IoT-based monitoring and control system also supports better farm management decisions, ensuring improved yield and more efficient use of resources. In simple terms, the Automated Vegetable Picker Machine blends robotics, AI and IoT to give farmers a practical, smart harvesting solution helping them save time, improve vegetable quality and move a step closer toward sustainable, technology-driven agriculture.
1. INTRODUCTION
In today’s agricultural landscape, the integration of automation and Internet of Things (IoT) is transforming traditional farming into a more efficient, data-driven and sustainable system. Among the many applications, vegetable harvesting remains one of the most challenging and labor-intensive stages. To address this, the proposed project introduces an Automated Vegetable Picker Machine, a solution that combines IoT and Artificial Intelligence (AI) to detect and harvest ripe Vegetables accurately and reliably. The core of this system lies in its advanced components. The core of this system lies in its advanced components. Real-time image capture with a digital camera, an RGB color sensor for analyzing vegetable ripeness and a 4-DOF robotic arm capable of performing precise harvesting tasks. An AI model, trained with platforms like Google Teachable Machine or TensorFlow, is deployed on the ESP32-CAM to dynamically classify ripeness levels based on Vegetable color and size. When a ripe vegetable is detected, the Node MCU microcontroller processes the data and sends signals to the robotic arm, guiding its movements for efficient picking.
Enhanced functionality comes from the system’s seamless IoT connectivity with a mobile application. Farmers receive real-time updates, notifications and insights into harvesting progress, making the process transparent and accessible. By automating vegetable selection and collection, this machine reduces dependency on manual labor, improves harvest consistency, minimizes post- harvest damage and helps optimize overall yield and resource utilization. The design also incorporates IoT-based monitoring and decision-making tools, enabling smarter farm management. This points toward a future of precision and sustainable farming, where technology reduces uncertainty and effort while maximizing output.
From a machine vision perspective, the system is structured into three main stages:
i. Receiving images – A wireless digital camera captures environmental images of the Vegetable and surrounding area.
ii. Processing images – Using OpenCV and AI algorithms, the captured images are analyzed to identify ripeness and measure distance with the assistance of ultrasonic sensors.
iii. Sending information – The processed data is transferred to the control unit, which directs motor movements to adjust the robotic arm and vehicle navigation.
Ultrasonic sensors assist in determining the distance between the robotic arm and the Vegetable, while DC motors and servo motors execute the required movements. This ensures that the robotic arm can effectively reach and harvest the target Vegetable without damage.
2. OBJECTIVES
Ø To build an automated machine that can pick ripe vegetables with care and accuracy.
Ø To use a camera and AI model to detect when vegetables are ripe.
Ø To design a robotic arm that can harvest vegetables without damaging them.
Ø To connect the system with a mobile app so farmers can monitor and control it in real time.
Ø To reduce the need for manual labor and save time during harvesting.
Ø To test the machine in field conditions and check its efficiency and usefulness.
3. METHODOLOGY
Image and Ripeness Detection: The ESP32-CAM captures live images of vegetables, while the color sensor continuously monitors color data to classify vegetable ripeness using an AI model deployed on the microcontrollers.
Signal Processing: When a ripe vegetable is detected, NodeMCU processes this information and generates control signals. These signals guide both the robotic arm and drive motors via Arduino Uno and the L298N driver.
Picking Operation: The servo motors precisely articulate the 4-DOF robotic arm to extend, position and operate the claw to gently pluck the vegetable without causing damage.
Mobility and Navigation: The 4-wheel drive chassis moves within orchard rows, avoiding obstacles and aligning with vegetable locations as instructed by the microcontroller logic and sensor feedback.
Control Interface: During development or in semi-autonomous modes, a joystick can be used to manually guide the arm and vehicle, allowing testing and adjustments.
IoT Integration: Data on harvesting progress, battery status and vegetable counts are sent via Node-MCU to a connected mobile application using Wi-Fi or Bluetooth, enabling farmers to remotely monitor and control operations.
4. COMPONENTS
Servo Motors: These actuate the joints of the robotic arm, allowing precise and smooth movements to reach and pick vegetables gently.
Microcontrollers (Node-MCU and Arduino Uno): Node-MCU manages IoT connectivity and processes sensor data, while Arduino Uno controls the motors and robotic arm operations.
Bluetooth Module: Enables wireless communication for control and data transfer between devices.
L298N Motor Driver: Controls the direction and speed of DC motors that move the chassis and robotic arm joints.
Color Sensor: Detects the ripeness of vegetables based on color analysis, feeding data to microcontrollers for real-time decision-making.
Robotic Arm (4-DOF): A four degrees-of-freedom arm provides the flexibility to maneuver and pluck Vegetables with minimal damage.
4-Wheel Drive Chassis: Allows the vehicle to navigate between orchard rows and position itself for Vegetable picking.
Joystick: Provides manual control for testing, calibration, or overriding autonomous operations.
5. WORKING PRINCIPLE
The Automated Vegetable Picker Machine operates through a coordinated process of sensing, intelligence and actuation to autonomously identify and harvest ripe vegetables. The system begins by capturing images of the vegetable-bearing plants using a camera module alongside a color sensor that detects specific color ranges corresponding to vegetable ripeness. These visual inputs are processed in real time by an AI model deployed on microcontrollers like Node-CU and Arduino Uno, which classify vegetables as ripe or unripe based on trained algorithms analyzing color, size and texture features. Upon detecting a ripe vegetable, the control unit calculates its location relative to the robotic arm. The machine’s 4 degrees-of-freedom robotic arm, powered by servo motors, then executes precise movements to maneuver to the target vegetable. An end-effector or gripper mimics human hand motions to gently grasp and separate the vegetable from the plant, minimizing bruising or damage. Meanwhile, the four-wheel drive chassis moves the vehicle along orchard rows under the direction of motor drivers controlled by the microcontroller, positioning the picker accurately for successive harvests. The system uses sensor feedback to avoid obstacles and adjust navigation dynamically.
Further enabling real-time monitoring and farmer interaction, the entire machine is connected via IoT to a mobile application, which provides live updates, alerts on battery and vegetable counts and options for remote control or manual override using a joystick when needed.
This seamless integration of computer vision, AI decision-making, precise robotic actuation and IoT connectivity allows the system to automate the Vegetable picking task efficiently and consistently, significantly reducing labor needs while improving harvesting quality and productivity in orchard environments.
6. TESTING
a. Functional Testing
Functional testing of the smart vegetable picker is performed using an Arduino UNO board, a Bluetooth module and an L298N motor driver. Following the circuit design, the Arduino UNO, Bluetooth module and L298N motor driver are connected. An Arduino sketch integrating necessary features such as motor control and Bluetooth connectivity is uploaded, including relevant libraries as needed. The Bluetooth module is paired with a smartphone or computer. Commands like start, stop and speed control are sent via Bluetooth to the Arduino. It is verified that the commands are properly received and executed by the system to control the vegetable picking mechanism.
b. Performance Testing
Performance testing assesses the smart vegetable picker’s operation focusing on the Arduino UNO, Bluetooth module and L298N motor driver. The system's ability to connect with the Bluetooth module, receive wireless commands and operate the motor driver to control the picker is examined. The Bluetooth connection is stable and functional within a 10-meter range, establishing that the vegetable picker has an effective communication range of about 10 meters.
7. RESULT AND DISCUSSION
The Automated Vegetable Picker Machine effectively combined an ESP32- CAM for image capture, a color sensor for ripeness detection and a 4-DOF robotic arm for harvesting. The AI model, trained with Google Teachable Machine/TensorFlow, successfully classified ripe vegetables with good accuracy, ensuring only mature vegetables were picked while minimizing damage. The Node MCU reliably processed data and controlled the robotic arm, enabling precise harvesting. IoT connectivity allowed farmers to monitor operations through a mobile app, offering real-time updates and better decision-making. Key benefits observed include reduced reliance on manual labor, greater efficiency, lower post-harvest losses and improved yield quality. Some challenges were noted, such as reduced accuracy under poor lighting, limited battery backup and slower picking speed compared to human labor. Overall, the system demonstrates that IoT and AI-based automation can address critical agricultural challenges, paving the way toward more efficient and sustainable farming practices.
Author Bios:
- B. Bhavadharani: Assistant Professor, Department of Agricultural Engineerin
- R. Kavinila: Assistant Professor, Department of Agricultural Engineering.
- G. Jacob Samson: UG Student, Department of Agricultural Engineering.
- R. Prakash: UG Student, Department of Agricultural Engineering.
Comments
Post a Comment