🌱 D455 Lawn Care System

Intel RealSense D455 autonomous lawn care system with ROS 2 Humble

Complete computer vision pipeline for grass detection, obstacle avoidance, and coverage path planning using depth and color cameras.

📁 Package Structure

~/lawn_care_ws/
├── src/d455_lawn_care/           # Main package directory
│   ├── d455_lawn_care/           # Python package
│   │   ├── grass_detector.py     # Grass detection node
│   │   └── obstacle_detector.py  # Obstacle detection node
│   ├── msg/                      # Custom message definitions
│   │   ├── GrassDetection.msg    # Grass analysis results
│   │   ├── Obstacle.msg          # Individual obstacle data
│   │   ├── ObstacleArray.msg     # Multiple obstacles
│   │   └── LawnStatus.msg        # Overall system status
│   ├── srv/                      # Service definitions
│   │   ├── CalibrateGrassDetection.srv  # Calibration service
│   │   └── SetMowingPattern.srv  # Pattern planning service
│   ├── launch/                   # Launch files
│   │   └── lawn_care_system.launch.py  # Complete system launch
│   ├── config/                   # Configuration files
│   │   └── lawn_care_params.yaml # Tuned parameters
│   └── package.xml              # Package dependencies

🚀 Quick Start

1. Build the Workspace

cd ~/lawn_care_ws
colcon build --packages-select d455_lawn_care
source install/setup.bash

2. Launch D455 Camera (already running)

# Your D455 is already streaming with optimal settings
ros2 topic list | grep d455

3. Launch Lawn Care System

source ~/lawn_care_ws/install/setup.bash
ros2 launch d455_lawn_care lawn_care_system.launch.py

4. Run Individual Nodes

# Grass detection
python3 ~/lawn_care_ws/src/d455_lawn_care/d455_lawn_care/grass_detector.py

# Obstacle detection  
python3 ~/lawn_care_ws/src/d455_lawn_care/d455_lawn_care/obstacle_detector.py

🌱 Features

Grass Detection Node

• Real-time grass analysis using HSV color filtering
• Height measurement using RGB-D depth data
• Coverage percentage calculation
• Cutting recommendations based on height thresholds
• Auto-calibration service for different grass types
• Debug visualizations with grass mask overlay

Obstacle Detection Node

• 3D obstacle detection using depth clustering
• Obstacle classification (tree, rock, furniture, etc.)
• Persistent tracking with unique IDs
• Safety zone calculation with danger levels
• RViz visualization with colored bounding boxes
• Configurable detection parameters

Custom Messages

• GrassDetection.msg - Complete grass analysis data
• Obstacle.msg - Individual obstacle properties
• ObstacleArray.msg - Multiple obstacles with statistics
• LawnStatus.msg - Overall system health and status

Services

• CalibrateGrassDetection.srv - Adaptive grass color calibration
• SetMowingPattern.srv - Autonomous mowing path planning

🔧 Configuration

All parameters are tuned for D455 lawn care in config/lawn_care_params.yaml:

Grass Detection Parameters

grass_detector:
  ros__parameters:
    # HSV color range (outdoor optimized)
    grass_color_hue_min: 40      # Green hue lower bound
    grass_color_hue_max: 80      # Green hue upper bound
    
    # Height thresholds
    min_grass_height: 0.02       # 2cm minimum
    cutting_threshold: 0.08      # 8cm cutting trigger
    
    # Performance
    detection_frequency: 10.0    # 10 Hz detection rate

Obstacle Detection Parameters

obstacle_detector:
  ros__parameters:
    # Detection ranges
    min_obstacle_height: 0.10    # 10cm minimum height
    max_detection_distance: 10.0 # 10m maximum range
    
    # Safety zones
    safety_buffer_distance: 0.5  # 50cm safety buffer
    danger_zone_distance: 1.0    # 1m danger zone

📊 Topics Published

From Grass Detector:

• /lawn_care/grass_detector/grass_detection - Grass analysis results
• /lawn_care/grass_detector/debug_image - Visualization with overlay
• /lawn_care/grass_detector/grass_mask - Binary grass mask

From Obstacle Detector:

• /lawn_care/obstacle_detector/obstacles - Detected obstacles array
• /lawn_care/obstacle_detector/obstacle_markers - RViz visualization
• /lawn_care/obstacle_detector/debug_image - Debug visualization

From D455 Camera:

• /d455/d455_camera/color/image_raw - RGB camera feed
• /d455/d455_camera/aligned_depth_to_color/image_raw - Aligned depth
• /d455/d455_camera/depth/color/points - Point cloud data

🎯 Current Status

✅ Package Built Successfully:

• ROS 2 workspace created and configured
• Custom messages and services defined
• Python nodes with OpenCV dependencies
• Launch files for system integration
• Configuration parameters optimized for D455

✅ D455 Camera Active:

• Streaming at 1280x720 @ 30fps
• Outdoor-optimized settings applied
• All depth processing filters active
• Perfect integration with ROS 2 topics

✅ Ready for Integration:

• Grass detection algorithms implemented
• Obstacle detection and tracking ready
• Calibration services available
• Debug visualizations enabled

🔄 Next Steps

1. Test Grass Detection:

   python3 ~/lawn_care_ws/src/d455_lawn_care/d455_lawn_care/grass_detector.py

2. Test Obstacle Detection:

   python3 ~/lawn_care_ws/src/d455_lawn_care/d455_lawn_care/obstacle_detector.py

3. Launch Complete System:

   ros2 launch d455_lawn_care lawn_care_system.launch.py

4. Calibrate for Your Grass:

   ros2 service call /lawn_care/grass_detector/calibrate_grass_detection d455_lawn_care/srv/CalibrateGrassDetection "{auto_calibrate: true}"

🛠️ Dependencies Included

• OpenCV for computer vision processing
• NumPy for numerical computations
• SciPy for advanced mathematical functions
• scikit-learn for machine learning features
• Matplotlib for data visualization
• cv_bridge for ROS-OpenCV integration
• tf2_ros for coordinate transformations

📝 Notes

• The package is built as a pure Python package for easier development
• All parameters are configurable via YAML files
• Debug visualizations help with tuning and debugging
• Custom messages provide structured data for lawn care applications
• Services enable dynamic reconfiguration and calibration