Fire Satellite Image Crawler

This tool fetches and processes satellite imagery data related to fire events using NASA's Earth API and Google Earth Engine.

Features

• Search for active fires using MODIS fire products
• Get high-resolution satellite imagery from Sentinel-2 or Landsat 8
• Apply specialized fire detection indices to enhance visualization
• Filter by date range, region, and fire detection confidence
• Automatic cloud cover filtering for best quality
• Batch download with progress tracking

Setup

Requirements

• Python 3.7 or higher
• Google Earth Engine account
• NASA API key

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/fire-satellite-image-crawler.git
   cd fire-satellite-image-crawler
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Setup environment variables: Create a .env file in the project root with:

   NASA_API_KEY=your_nasa_api_key
   PROJECT-ID=your_gee_project_id
   

Google Earth Engine Authentication

Before running the script for the first time, you need to authenticate with Google Earth Engine:

Usage

Basic Usage

Run the main script:

python main.py --start_date 2023-01-01 --end_date 2023-01-31 --satellite sentinel

Custom Region

Search in a specific region:

python main.py --start_date 2023-07-01 --end_date 2023-08-01 --region -120.5,37.7

Advanced Usage

from fire_crawler import FireSatelliteImageCrawler

# Initialize the crawler
crawler = FireSatelliteImageCrawler(mpc_api_key="your_api_key_here")

# Search for fires in California during summer 2023
images = crawler.search_and_download_fire_images(
    start_date="2023-06-01",
    end_date="2023-09-01",
    bbox=[-124.4, 32.5, -114.1, 42.0],  # California bounding box
    satellite="sentinel",
    confidence="high",
    output_dir="./california_fires"
)

Parameters

• start_date/end_date: Date range for fire search in 'YYYY-MM-DD' format
• satellite: 'sentinel' (default) or 'landsat8'
• confidence: Fire detection confidence - 'high' (default), 'nominal', or 'low'
• region: Geographic region specified as 'lon,lat' (creates a 100km box around this point)
• max_cloud_cover: Maximum cloud cover percentage (default: 20)
• days_after_fire: Number of days after fire detection to look for imagery (default: 7)
• visualization: Visualization style - 'rgb' (natural color) or 'fire' (fire-enhanced)
• output_dir: Directory to save downloaded images (default: ./output)

Microsoft Planetary Computer

This tool uses the Microsoft Planetary Computer, which provides a vast catalog of environmental data including:

• Sentinel-2 multispectral imagery
• Landsat 8 imagery
• MODIS fire products
• VIIRS fire detections
• And many more datasets

To learn more, visit the Microsoft Planetary Computer website.

Troubleshooting

Common Errors

1. No fire points found

This typically happens when:

• No fires occurred in the specified region/time range
• The date range is too narrow or in an inactive fire season
• The region is not correctly specified

Solutions:

• Try a wider date range
• Try a different region
• Lower the confidence level with --confidence nominal

2. No images downloaded

This can occur when:

• No suitable satellite imagery exists for the detected fire points
• Cloud cover is too high over all fire locations
• The satellite didn't capture the area during the specified time window

Solutions:

• Increase --days_after_fire to look for imagery in a wider time window
• Increase --max_cloud_cover to accept images with more cloud coverage
• Try a different satellite with --satellite landsat8

3. Authentication errors

If you see authentication errors:

• Check that your Microsoft Planetary Computer API key is valid
• Make sure the API key is correctly specified in config.py or via --mpc_api_key

Examples

Find Recent Fires with Custom Parameters

python main.py \
  --start_date 2023-01-01 \
  --end_date 2023-02-01 \
  --region -120.5,37.7 \
  --satellite sentinel \
  --max_cloud_cover 30 \
  --days_after_fire 14 \
  --visualization fire