Skip to content
BranchesTags

Repository files navigation

| GEE AutoLab - Automated Remote Sensing Research Assistant

License: MIT Python 3.9+ GEE

GEE AutoLab is a fully automated remote sensing analysis agent built on Google Earth Engine. It transforms a simple natural language query (time + location + task) into a complete scientific workflow: data retrieval | GEE script generation | index/classification/change detection analysis | GeoTIFF/CSV export | experiment report generation.

Features

| Zero-Code Operation

  • Natural Language Input: Describe your analysis in plain Chinese or English
  • Auto Parameter Filling: Missing parameters are automatically filled with intelligent defaults
  • One-Click Execution: From query to full scientific report in one command

Comprehensive Data Source Support

  • Official GEE Datasets: Full support for 50+ public datasets including Sentinel, Landsat, MODIS, VIIRS, GPM, SRTM, ERA5, etc.
  • SAT-IO Community Catalog: Extended support for community-contributed datasets
  • Intelligent Auto-Selection: Automatically selects the optimal dataset based on task type and resolution requirements
  • Custom Dataset Support: Users can specify any GEE dataset by ID

| 12 Analysis Task Categories

  1. Vegetation Indices - NDVI, EVI, NDWI, MNDWI, NDBI, NDMI, SAVI, GCI, RECI, NBR, PSRI, SIPI, ARVI, GNDVI, NDRE, MSAVI (16+ indices)
  2. Water & Wetland Analysis - Water extraction, area statistics, wetland classification, shoreline change
  3. Land Cover Classification - Random Forest, CART, SVM, Gradient Boosted Trees with accuracy assessment
  4. Change Detection - LandTrendr, CCDC, bi-temporal differencing, spectral angle, PCA change detection
  5. Time Series Analysis - Annual/monthly composites, trend analysis, anomaly detection, phenology parameters
  6. Urban & Nightlight - Urban expansion, impervious surface, NPP-VIIRS analysis, urban heat island
  7. Forest & Carbon - Forest cover, disturbance detection, biomass estimation, carbon stock calculation
  8. Disaster Monitoring - Fire detection, burn area, flood mapping, drought monitoring, landslide detection
  9. Terrain & DEM - Elevation, slope/aspect, watershed delineation, topographic indices
  10. Meteorology & Climate - Precipitation, temperature, evapotranspiration, drought indices
  11. SAR Analysis - Sentinel-1 GRD, flood detection, ship detection, soil moisture
  12. Statistics & Export - Area statistics, zonal statistics, CSV export, chart generation

| Automatic Export

  • GeoTIFF - Cloud-optimized raster export
  • CSV - Statistical tables and time series data
  • Shapefile/GeoJSON - Vector analysis results
  • PNG/JPG - Visualization thumbnails
  • Cloud Optimized GeoTIFF - For efficient web delivery

| Scientific Report Generation

  • Research area overview
  • Data source documentation
  • Methodology description
  • Parameter configuration record
  • Result statistics and analysis
  • Accuracy assessment (for classification tasks)
  • Charts and visualizations
  • Limitations and improvement suggestions

| GEE Official Docs Integration

GEE AutoLab is built with reference to the complete GEE documentation ecosystem:

| Quick Start

Prerequisites

  • Python 3.9+
  • Google Earth Engine account (sign up)
  • Internet connection

Installation

`�ash

Clone the repository

git clone https://github.com/xingguangYan/GEE-AutoLab.git cd GEE-AutoLab

Install dependencies

pip install -r requirements.txt

Authenticate with GEE

earthengine authenticate

Verify installation

python scripts/check_gee.py `

Usage Examples

Example 1: Vegetation Index Calculation �ash python scripts/run_pipeline.py "在武汉市，2023年，计算NDVI和EVI" --output result_wuhan

Example 2: Land Cover Classification �ash python scripts/run_pipeline.py "地点：黑龙江省 时间：2022 任务：土地覆盖分类 分类器：Random Forest" --output result_heilongjiang

Example 3: Change Detection �ash python scripts/run_pipeline.py "云南省，2018-2023年，森林变化检测" --output result_yunnan_forest

Example 4: Time Series Analysis �ash python scripts/run_pipeline.py "长江流域，2000-2023年，NDVI时间序列趋势分析" --output result_yangtze_timeseries

Quick Modes

`�ash

Quiet mode - suppress detailed progress

python scripts/run_pipeline.py "||" --output output_dir -q

Generate GEE code only (no execution)

python scripts/run_pipeline.py "||" --output output_dir --code-only

Generate report only from existing results

python scripts/run_pipeline.py "||" --output output_dir --report-only `

Architecture

User Input (Natural Language) | | |||||||||| | task_parser | Parses location, time, task, parameters |||||||||| | | |||||||||| | generate_gee | Generates GEE JavaScript/Python code | | Indices analysis | | Classification | + data_sources | Change detection | | Time series |||||||||| | | |||||||||| | run_pipeline | Orchestrates full pipeline execution |||||||||| | GEE Code (.js) Task Execution (optional) | | |||||||||| | generate_report | Generates MD/HTML report |||||||||| | | Final Output: | GeoTIFF files | CSV statistics | Accuracy assessment | Charts/Visualizations | Experiment report (MD/HTML)

| Project Structure

GEE-AutoLab/ | SKILL.md # Skill definition (Codex) | README.md # English documentation | README_CN.md # Chinese documentation | LICENSE # MIT License | CITATION.cff # Citation information | requirements.txt # Python dependencies | scripts/ | | run_pipeline.py # Main entry point | | task_parser.py # Natural language parser | | generate_gee.py # GEE code generator | | generate_report.py # Report generator | | check_gee.py # Environment checker | templates/ | | report.md # Report template | references/ | | data_sources.md # Dataset reference (50+ datasets) | examples/ | ndvi_analysis.md # NDVI analysis example | classification.md # Land cover classification example | change_detection.md # Change detection example | timeseries.md # Time series analysis example

| Supported Data Sources

Optical Data

Dataset ID Resolution
Sentinel-2 MSI SR COPERNICUS/S2_SR 10m
Sentinel-2 MSI TOA COPERNICUS/S2 10m
Sentinel-2 MSI SR (Harmonized) COPERNICUS/S2_SR_HARMONIZED 10m
Landsat 8-9 OLI/TIRS SR LANDSAT/LC08/C02/T1_L2 30m
Landsat 7 ETM+ SR LANDSAT/LE07/C02/T1_L2 30m
Landsat 4-5 TM SR LANDSAT/LT05/C02/T1_L2 30m
Landsat Collection 1 SR LANDSAT/LC08/C01/T1_SR 30m
MODIS MOD09GA MODIS/006/MOD09GA 500m
MODIS MOD09Q1 MODIS/006/MOD09Q1 250m
MODIS MOD13Q1 MODIS/006/MOD13Q1 250m

SAR Data

Dataset ID Resolution
Sentinel-1 GRD COPERNICUS/S1_GRD 10m
ALOS PALSAR JAXA/ALOS/PALSAR/YEARLY/SAR 25m

Climate & Weather

Dataset ID Resolution
GPM IMERG NASA/GPM_L3_IMERG_V06 10km
ERA5 Monthly ECMWF/ERA5_LAND/MONTHLY 11km
CHIRPS Daily UCSB-CHG/CHIRPS/DAILY 5km
MODIS LST MODIS/006/MOD11A2 1km
GRIDMET IDAHO_EPSCOR/GRIDMET 4km

Land Cover

Dataset ID Resolution
ESA WorldCover ESA/WorldCover/v200 10m
Hansen Global Forest UMD/hansen/global_forest_change_v1_11 30m
JRC Global Surface Water JRC/GSW1_4/GlobalSurfaceWater 30m
MODIS Land Cover MODIS/006/MCD12Q1 500m
Copernicus CORINE COPERNICUS/CORINE/V20_100 100m

Terrain

Dataset ID Resolution
SRTM Digital Elevation USGS/SRTMGL1_003 30m
Copernicus DEM COPERNICUS/DEM/GLO30 30m
NASADEM NASA/NASADEM_HGT/001 30m

Fire & Thermal

Dataset ID Resolution
MODIS Fire/Months MODIS/006/MOD14A2 1km
VIIRS S-NPP Fire VIIRS/VNP14IMG 375m
MODIS Thermal MODIS/006/MOD11A2 1km

Nightlight

Dataset ID Resolution
NPP-VIIRS VNP46A4 VIIRS/VNP46A4 500m
DMSP-OLS NOAA/DMSP-OLS/NIGHTTIME_LIGHTS 1km

| Example Output

NDVI Analysis Report

  • Location: Wuhan, China
  • Data: Sentinel-2 SR (10m resolution)
  • Period: 2023-01-01 to 2023-12-31
  • Indices: NDVI, EVI, MNDWI
  • Results:
    • Mean NDVI: 0.45 (range: -0.12 to 0.89)
    • Vegetation coverage: 62.3% of study area
    • Water area: 8.7% (derived from MNDWI)
  • Exports: GeoTIFF x 3, CSV x 1

Land Cover Classification (Random Forest)

  • Overall Accuracy: 92.4%
  • Kappa Coefficient: 0.89
  • Classes: 6 land cover types
  • Exports: Classified GeoTIFF + Accuracy table (CSV)

| Customization

Adding Custom Datasets

Edit eferences/data_sources.md to add dataset entries: `yaml

  • name: My Custom Dataset id: PATH/TO/DATASET resolution: 30 type: optical tasks: [classification, indices] `

Creating Custom Templates

Edit emplates/report.md to customize report format and content.

| Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

| License

Distributed under the MIT License. See LICENSE for more information.

| Citation

�ibtex @software{gee_autolab, author = {{GEE AutoLab Contributors}}, title = {GEE AutoLab: Automated Remote Sensing Research Assistant}, year = {2025}, url = {https://github.com/xingguangYan/GEE-AutoLab} }

| Acknowledgments

  • Google Earth Engine for providing the powerful cloud computing platform
  • SAT-IO community for the extended dataset catalog
  • ESA, NASA, USGS, JAXA for open Earth observation data
  • CSDN Blog Reference

About

Fully automated remote sensing research pipeline for Google Earth Engine. One sentence in - full analysis, GeoTIFF/CSV exports, and experiment report out.

blog.csdn.net/qq_31988139

Resources

Readme
MIT license
Activity

Stars

4 stars

Watchers

0 watching

Forks

1 fork
Report repository

Releases

Packages

Contributors

Languages