Rasteric
Rasteric is a Python geospatial library designed for raster preprocessing, analysis, visualization, and modeling. It simplifies common GIS workflows with easy-to-use functions built on top of Rasterio and GeoPandas.
Installation
pip install rasteric
Dependencies are installed automatically:
- rasterio, geopandas, shapely
- numpy, pandas, matplotlib, scipy
- scikit-learn, scikit-image
- rasterstats, Pillow
Requirements
Supported Data Formats
- Raster: GeoTIFF, TIFF
- Vector: Shapefiles, GeoJSON
- Tabular: CSV with spatial attributes (latitude/longitude)
Typical Use Cases
- Satellite image preprocessing (Sentinel-2, Landsat)
- NDVI and vegetation analysis
- Land-use / land-cover mapping
- Agricultural monitoring
- Raster-vector data extraction
- Geospatial machine learning data preparation
- Time-series crop monitoring
- Change detection
- Pan-sharpening (spatial resolution enhancement)
Quick Start
from rasteric import (plot, clip, ndvi, evi, stack, extract, mosaic, calc,
change_detect, timeseries, cloudmask, classify, sample)
# Visualize
plot('sentinel2.tif', bands=(4, 3, 2), brightness_factor=4)
# Vegetation indices
ndvi('sentinel2.tif', 'ndvi.tif', red_band=3, nir_band=4)
evi('sentinel2.tif', 'evi.tif', nir_band=4, red_band=3, blue_band=2)
# Mosaic tiles
mosaic('tiles/', 'merged.tif')
# Band math
calc('(B8 - B4) / (B8 + B4)', inputs={'B8': 'nir.tif', 'B4': 'red.tif'}, output_file='ndvi.tif')
# Change detection
result = change_detect('ndvi_jan.tif', 'ndvi_jul.tif', 'change.tif', threshold=0.2)
# Pan-sharpen: boost RGB from 5m to 1m using panchromatic band
rasteric.pansharpen('rgb_5m.tif', 'pan_1m.tif', 'rgb_1m.tif', method='gram-schmidt')
# Time-series
df = timeseries('ndvi_folder/', output_csv='ndvi_ts.csv')
# Cloud masking (Sentinel-2)
cloudmask('sentinel2.tif', 'cloud_free.tif', method='scl')
# Classification
classify('image.tif', 'classes.tif', method='kmeans', n_clusters=5)
# ML training data
df = sample('sentinel2.tif', 'training.shp', 'samples.csv', label_column='class')
Functions
Vegetation Indices
| Function |
Description |
ndvi(raster, out, red, nir) |
Normalized Difference Vegetation Index |
evi(raster, out, nir, red, blue) |
Enhanced Vegetation Index |
savi(raster, out, nir, red, L) |
Soil-Adjusted Vegetation Index |
ndwi(raster, out, green, swir) |
Normalized Difference Water Index |
gndvi(raster, out, nir, green) |
Green NDVI (chlorophyll sensitivity) |
ndmi(raster, out, nir, swir) |
Normalized Difference Moisture Index |
Data Handling
| Function |
Description |
stack(folder, out) |
Stack multiple rasters into multi-band file |
mosaic(folder, out, method) |
Merge raster tiles into one raster |
mergecsv(path, out) |
Merge CSV files in a directory |
clip(raster, shp, out) |
Clip raster by shapefile |
reproject(raster, out, crs) |
Reproject to different CRS |
resample(raster, out, factor) |
Resample to new resolution |
Raster Calculator
| Function |
Description |
calc(expr, inputs, out) |
Band math with named inputs |
bandmath(raster, equation, out) |
Band math with B1, B2, ... |
Analysis
| Function |
Description |
zonalstats(raster, vector, stats) |
Zonal statistics for polygons |
stats(raster) |
Basic statistics (min, max, mean, std) |
fillnodata(raster, out) |
Fill NoData via interpolation |
Time-Series & Change Detection
| Function |
Description |
timeseries(folder, out_csv, band, stat) |
Analyze raster time-series |
change_detect(img1, img2, out, method, threshold) |
Detect changes between dates |
Cloud Masking
| Function |
Description |
cloudmask(raster, out, method) |
Sentinel-2 cloud masking (QA60 or SCL) |
Image Classification
| Function |
Description |
classify(raster, out, method, n_clusters) |
K-Means or Random Forest classification |
texture(raster, out, band, properties) |
GLCM texture features |
ML Preparation
| Function |
Description |
sample(raster, shp, out_csv, label) |
Extract training samples from raster |
train_test_split(csv, test_size) |
Split samples into train/test sets |
Extraction
| Function |
Description |
extract(data, shp, out_csv) |
Extract raster values at vector locations |
align_to_shp(tif, shp, out) |
Reproject raster to match shapefile CRS |
Visualization
| Function |
Description |
plot(file, bands, cmap, title, ax, brightness) |
Display raster with band selection |
contour(file) |
Overlay contour lines |
hist(file, bin, title) |
Histogram of pixel values |
thumbnail(raster, out_png, bands, size) |
Quick PNG thumbnail |
Conversion
| Function |
Description |
convras(raster, out_shp) |
Raster to vector polygons |
Pan-sharpening
| Function |
Description |
pansharpen(ms_raster, pan_raster, output, method) |
Increase MS resolution using pan band |
import rasteric
# Pan-sharpen RGB (5m) using panchromatic (1m) → output 1m
rasteric.pansharpen('rgb_5m.tif', 'pan_1m.tif', 'rgb_1m.tif', method='gram-schmidt')
Methods:
brovey — Brovey Transform, sharpest spatial detailihs — IHS (Intensity-Hue-Saturation), good color preservationgram-schmidt — best spectral integrity (recommended for classification)simple_mean — simple average fallback
Utilities
| Function |
Description |
convpath(path) |
Cross-platform path conversion |
bandnames(raster, names) |
Update band descriptions |
haversine(lon1, lat1, lon2, lat2) |
Great-circle distance |
Sentinel-2
from rasteric import process_sen2
# Full pipeline: JP2→TIFF, resample, stack, cloud removal
process_sen2('source/', 'output/')
Contributing
License
MIT License
