Run the Mining Workflow from Python¶

This guide is based on examples/package_usage.ipynb and examples/Mining.ipynb.

Goal¶

Find a smaller secondary image inside a larger reference raster from Python.

Inputs¶

You need:

• a reference raster path
• a secondary image path
• optionally, rank
• optionally, fdecimation

Minimal call¶

import gefolki

xmin, xmax, ymin, ymax, reference_final = gefolki.mining(
    "datasets/S1_Jacksonville_GEE.tif",
    "datasets/JacksonvilleNavalAirStation_sandiaKu.png",
)

Tuned call¶

import gefolki

xmin, xmax, ymin, ymax, reference_final = gefolki.mining(
    file_path_reference="datasets/S1_Jacksonville_GEE.tif",
    file_path_secondary="datasets/JacksonvilleNavalAirStation_sandiaKu.png",
    rank=4,
    fdecimation=8,
)

What happens¶

The mining workflow:

1. loads the reference raster and secondary image
2. computes a coarse search on downsampled data
3. refines the match in a cropped reference region
4. returns the bounding coordinates and extracted reference patch
5. displays diagnostic Matplotlib figures

When to open the notebook¶

If you want to inspect the intermediate steps, open:

• examples/Mining.ipynb

That notebook breaks the algorithm into the same stages used by gefolki.mining(...).