EcoCanopy

Processing pipeline for ECOSENSE camera data that extracts vegetation metrics (leaf angles, reflectance, GCC) from video frames.

Installation

cd /mnt/data/lk1167/projects/EcoCanopy
pip install -e .

Pipeline Stages

1. Preprocessing: Extract timestamps, classify image modes (RGB/NIR)
2. Reflectance: Compute mean/median reflectance for NIR images
3. GCC: Compute Green Chromatic Coordinate for RGB images
4. Leaf Angles: Run AngleCam model inference
5. Upload: Store results in SQLite database
6. Dashboard: Visualize data via Dash frontend

Usage

Process a single camera

python -m src.cli preprocess -c G5Bullet_07
python -m src.cli gcc -c G5Bullet_07
python -m src.cli reflectance -c G5Bullet_07
python -m src.cli angles -c G5Bullet_07
python -m src.cli upload -c G5Bullet_07

Run complete pipeline

python -m src.cli run-all -c G5Bullet_07

Run for all configured cameras

python -m src.cli run-all

Launch dashboard

python -m src.cli dashboard

Check status

python -m src.cli status
python -m src.cli cameras

Configuration

Edit config/config.yaml to configure:

• Raw data paths
• Output paths
• Camera list
• Processing parameters

Data Structure

Input:

raw_videos/{camera_id}/{year}/{month}/{day}/
├── {camera_id}_output_video.mp4
└── {camera_id}_frame_data.csv

Output:

processed/{year}/cameras/{camera_id}/{month}/{day}/
└── {camera_id}_frames.csv

Tests

pytest tests/ -v