The AI-based mechanical weeding robot Nindamani uses artificial intelligence to automatically detect and segment crop weeds. The entire robot module is based on ROS2. Nindamani can be used for automatic weeding at any early stage of the crop. This robot is a proof of concept and prototype for agricultural robotics.
This is just the beginning of agricultural robot technology!
➡ Purpose:
To develop an autonomous weeding robot to address issues such as excessive herbicide use, harmful chemicals, and labor shortages in farmland.
➡ Vision:
Achieve herbicide-free farming and increase farmers’ yields.
➡ Project Impact:
• The robot will reduce the use of herbicides and pesticides
• Improve soil health and increase yield
• Reduce chemically grown food, which will improve human health
• Save farmers time and money
➡ Success Measures:
The robot’s performance is measurable if it can clear more than 75% of weeds from the entire agricultural field
➡ Final Agri-Robot Features:
• Application: Automatic weeder
• Four-wheel drive, independently steered robot
Weeding robot arm
• Uses artificial intelligence (AI) for weed detection
• Battery powered
• Navigation with camera and GPS
• Mobile operation
• Works for 8-10 hours
Materials used for the entire prototype:
Hardware:
-
NVIDIA Jetson NANO Development Kit *1
-
Stepper Motors *3
-
Stepper Drivers *3
-
Servo Motor *1
-
Arduino Mega 2560 & Genuino Mega 2560 *1
-
Battery *1
-
Raspberry Pi Camera Module V2
Software:
-
ROS
-
TensorFlow
Mechanism:
-
3D Printer
-
CNC
Execution Steps:
1. Flash NVIDIA JetPack
2. Install TensorFlow (tutorial available on NVIDIA’s official website)
3. ROS2 (Dashing Diademata)
4. Connect Arduino
5. Install OpenCV 3.4.4
6. Pay attention to install WIFI
7. Generate ROS2 Workspace
8. Mirror Mask R-CNN GitHub Repository
9. Download pre-trained model weights
10. Compile nindamani_ws
11. Implement stepper motor library on Arduino
12. Start nindamani robot
Check out the complete video introduction:
More Jetson development projects:
Don’t ask me how many stars there are, Jetson NANO tells you there are many….
Using Jetson NANO, create a Christmas tree that can save bees
Jetson NANO Application Case Showcase | Creating Smart Commercial Stores
Jetson NANO Application Case Showcase | Drones’ positioning flight in dark scenes
[Jetson Development Project Showcase] Create a smart pothole detector for roads using Jetson NANO and TensorRT
[Jetson Development Project Showcase] Create a smart delivery cabinet prototype with Jetson NANO
[Jetson Development Project Showcase] Dare to play rock-paper-scissors with NVIDIA JetBot?
[Jetson Development Project Showcase] Implement facial recognition attendance with Jetson NANO
[Jetson Development Project Showcase] How to use the hardware-accelerated version of FFMPEG for Jetson Nano
