Install
openclaw skills install @ruiduobao/geoskill-sar-forest-biomassSAR 森林生物量估算:由后向散射 σ⁰ 用线性/饱和经验模型反演森林地上生物量 AGB (t/ha),支持 C/L 波段与地面样本标定
openclaw skills install @ruiduobao/geoskill-sar-forest-biomassEstimates forest above-ground biomass (AGB, t/ha) from the SAR backscatter coefficient σ⁰. Two empirical relationships are implemented:
--calibration accepts a ground-sample CSV (columns sigma0,agb) for coefficient fitting: least squares for the linear model and nonlinear least squares (scipy curve_fit) for the saturation model. Without calibration, built-in default coefficients per band (C / L) are used; the L band has stronger penetration and a higher saturation biomass.
pip install 'numpy' 'rasterio' 'scipy' 'scikit-learn'
python geoskill-sar-forest-biomass.py --bbox 110.0 22.0 111.0 23.0 --band l --synthetic --output-dir ./out
python geoskill-sar-forest-biomass.py --bbox 110.0 22.0 111.0 23.0 --band c --model linear --synthetic --output-dir ./out
python geoskill-sar-forest-biomass.py --input sigma0_db.tif --band c --model linear --calibration samples.csv --output-dir ./out
python geoskill-sar-forest-biomass.py --input sigma0_db.tif --band l --model saturation --calibration samples.csv --output-dir ./out
| File | Format | Description |
|---|---|---|
forest_biomass.tif | GeoTIFF | Retrieved AGB (t/ha) |
biomass_report.json | JSON | Model coefficients, statistics, synthetic validation (RMSE/correlation) |
output-manifest.json | JSON | Run manifest |
Local σ⁰ (dB) GeoTIFF plus optional calibration CSV, or a simulated scene generated by forward modeling of an AGB field via --synthetic.
--synthetic mode is fully network-free.MIT
从 SAR 后向散射系数 σ⁰ 估算森林地上生物量 AGB(above-ground biomass, t/ha)。 实现两类经验关系:
支持 --calibration 提供地面样本 CSV(列 sigma0,agb)做系数拟合:线性用最小
二乘,饱和用非线性最小二乘(scipy curve_fit)。无标定时使用按波段(C / L)内置的
默认系数,L 波段穿透性强、饱和生物量更高。
pip install 'numpy' 'rasterio' 'scipy' 'scikit-learn'
python geoskill-sar-forest-biomass.py --bbox 110.0 22.0 111.0 23.0 --band l --synthetic --output-dir ./out
python geoskill-sar-forest-biomass.py --bbox 110.0 22.0 111.0 23.0 --band c --model linear --synthetic --output-dir ./out
python geoskill-sar-forest-biomass.py --input sigma0_db.tif --band c --model linear --calibration samples.csv --output-dir ./out
python geoskill-sar-forest-biomass.py --input sigma0_db.tif --band l --model saturation --calibration samples.csv --output-dir ./out
| 文件 | 格式 | 说明 |
|---|---|---|
forest_biomass.tif | GeoTIFF | 反演 AGB (t/ha) |
biomass_report.json | JSON | 模型系数、统计、合成验证 (RMSE/相关) |
output-manifest.json | JSON | 运行清单 |
本地 σ⁰ (dB) GeoTIFF + 可选标定 CSV,或 --synthetic AGB 场正演的模拟场景。
--synthetic 模式完全无网络。MIT