安全
高分辨率作物识别叠加申报地块做差异检测，核查补贴合规性。Verifies subsidy compliance by overlaying high-resolution crop classification on declared parcels for difference detection.
高分辨率作物识别叠加申报地块做差异检测，核查补贴合规性。Verifies subsidy compliance by overlaying high-resolution crop classification on declared parcels for difference detection.
Uses an NDVI raster derived from remote sensing imagery to classify crop/non-crop by threshold, overlays it with the vector parcels declared for subsidy, computes the "remotely sensed crop fraction" per parcel, and compares it with the "declared crop fraction": parcels exceeding the tolerance are flagged as suspected violations (over-declared / under-declared). This is a simplified implementation of the area-consistency check used in the EU CAP "Checks by Monitoring" and in national cropland subsidy verification.
Two verification methods:
area-diff (default): flagged when |measured fraction − declared fraction| > tolerance;class-match: binarizes declared/measured fractions into "crop/non-crop" at the 0.5 boundary, flagged when classes do not match.Data quality rules: NoData pixels are excluded from statistics; parcels with no valid pixels inside (or parcels smaller than the pixel and thus not burned in) are recorded as no-coverage and are not flagged as violations; NDVI values are validated to [-1,1]; bbox does not support crossing the 180° meridian.
pip install 'numpy' 'rasterio' 'geopandas' 'shapely'
python geoskill-agriculture-subsidy-verification.py --bbox 116.0 39.0 117.0 40.0 [other parameters]
python geoskill-agriculture-subsidy-verification.py --bbox 116.0 39.0 117.0 40.0 --synthetic --output-dir ./out
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --output-dir ./out
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --method class-match --tolerance 0.1
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --threshold 0.35
python geoskill-agriculture-subsidy-verification.py --input ndvi_utm.tif --parcels parcels_utm.gpkg --output-dir ./out
| File | Format | Description |
|---|---|---|
crop_mask.tif | GeoTIFF | Crop mask (1=crop, 0=non-crop, nodata in NoData areas) |
parcel_grid.tif | GeoTIFF | Parcel ID raster (background 0, parcels start from 1) |
verification_report.json | JSON | Per-parcel verification records and statistics (primary output) |
flagged_parcels.geojson | GeoJSON | Vector of suspected violation parcels |
output-manifest.json | JSON | Run manifest |
Local high-resolution NDVI GeoTIFF (EPSG:4326; projected coordinates are automatically reprojected) + declared parcel vectors (GeoJSON/GPKG/Shapefile, must contain parcel_id and declared_crop_frac columns, with declared fraction in [0,1]); --synthetic mode simulates offline without network.
--synthetic mode requires no network at all.MIT
用遥感影像派生的 NDVI 栅格做作物/非作物阈值分类，与申报补贴的地块矢量叠加， 逐地块计算"遥感实测作物占比"，与"申报作物占比"比较，超过容差即标记为疑似违规 （虚报 over-declared / 少报 under-declared）。这是欧盟 CAP "Checks by Monitoring" 与各国耕地补贴核查中面积一致性检查的简化实现。
两种核查方法：
area-diff（默认）：|实测占比 − 申报占比| > tolerance 即标记；class-match：以 0.5 为界把申报/实测二值化为"作物/非作物"，类别不一致即标记。数据质量规则：NoData 像元不参与统计；地块内无有效像元（或地块小于像元未被烧录）
记为 no-coverage，不判定违规；NDVI 值域校验 [-1,1]；bbox 不支持跨 180° 经线。
pip install 'numpy' 'rasterio' 'geopandas' 'shapely'
python geoskill-agriculture-subsidy-verification.py --bbox 116.0 39.0 117.0 40.0 [其他参数]
python geoskill-agriculture-subsidy-verification.py --bbox 116.0 39.0 117.0 40.0 --synthetic --output-dir ./out
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --output-dir ./out
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --method class-match --tolerance 0.1
python geoskill-agriculture-subsidy-verification.py --input ndvi.tif --parcels parcels.geojson --threshold 0.35
python geoskill-agriculture-subsidy-verification.py --input ndvi_utm.tif --parcels parcels_utm.gpkg --output-dir ./out
| 文件 | 格式 | 说明 |
|---|---|---|
crop_mask.tif | GeoTIFF | 作物掩膜（1=作物，0=非作物，NoData 区域为 nodata） |
parcel_grid.tif | GeoTIFF | 地块编号栅格（背景 0，地块从 1 起） |
verification_report.json | JSON | 逐地块核查记录与统计（主产物） |
flagged_parcels.geojson | GeoJSON | 疑似违规地块矢量 |
output-manifest.json | JSON | 运行清单 |
本地高分辨率 NDVI GeoTIFF（EPSG:4326；投影坐标自动重投影）+ 申报地块矢量
（GeoJSON/GPKG/Shapefile，需含 parcel_id 与 declared_crop_frac 列，
申报占比取值 [0,1]）；--synthetic 模式离线模拟，无需网络。
--synthetic 模式完全无网络。MIT
Detect and classify cropland changes from before/after NDVI rasters. Identifies suspected construction, water, forest, and bare soil changes and generates compliance investigation materials. Use when the user wants to check farmland for unauthorized changes, monitor cropland conversion, or generate compliance reports.
冠层峰值检测加分水岭分割统计株数，结合经验模型估算产量。Counts plants via canopy peak detection plus watershed segmentation and estimates yield with an empirical model.
Identify major crop types from multi-temporal optical/SAR imagery using phenological features. Produces pixel/field-level classification, area statistics, and confidence maps. Use when mapping crop distributions, estimating planted areas, or generating agricultural intelligence from remote sensing data.
融合 NDVI/NDRE/LST 多指数健康评分、历史偏差异常检测与空间聚类，输出作物健康等级栅格。Fuses NDVI/NDRE/LST into a health score with historical anomaly detection and spatial clustering.
融合 CWSI 水分胁迫、红边叶绿素与 SAR 含水量，检测作物胁迫等级。Detects crop stress by fusing CWSI water stress, red-edge chlorophyll and SAR water content.
基于红边异常、热红外温度、纹理变化与多时相早期胁迫检测，识别疑似病虫害区域。Detects suspected pest/disease areas from red-edge anomaly, thermal, texture and multi-temporal early stress.