Mosaic & merge¶

Combine several rasters into one:

• merge_rasters(src, dst) — mosaic many (overlapping or adjacent) rasters into a single raster covering their union.
• stack_bands(files, path=...) — stack several single-band rasters into one multi-band raster (e.g. assembling per-band Sentinel files into one image).

Setup¶

In [1]:

%matplotlib inline

import tempfile
from pathlib import Path

import numpy as np

DATA = Path('../../../examples/data')
WORK = Path(tempfile.mkdtemp(prefix='pyramids-ops-'))
DATA.is_dir(), WORK.is_dir()

%matplotlib inline import tempfile from pathlib import Path import numpy as np DATA = Path('../../../examples/data') WORK = Path(tempfile.mkdtemp(prefix='pyramids-ops-')) DATA.is_dir(), WORK.is_dir()

Out[1]:

(True, True)

In [2]:

from pyramids.dataset import Dataset
from pyramids.dataset.merge import merge_rasters, stack_bands

tiles = [DATA / 'crop_aligned_folder' / f'{i}.tif' for i in range(3)]
[Dataset.read_file(t).shape for t in tiles]

from pyramids.dataset import Dataset from pyramids.dataset.merge import merge_rasters, stack_bands tiles = [DATA / 'crop_aligned_folder' / f'{i}.tif' for i in range(3)] [Dataset.read_file(t).shape for t in tiles]

2026-07-11 14:41:46 | INFO | pyramids.base.config | Logging is configured.

Out[2]:

[(1, 24, 29), (1, 24, 29), (1, 24, 29)]

In [3]:

# Each input tile is a single-band 24x29 raster on a shared grid.
# `Dataset.plot(band=0)` renders a band inline (it returns a cleopatra ArrayGlyph).
Dataset.read_file(tiles[0]).plot(band=0)

# Each input tile is a single-band 24x29 raster on a shared grid. # `Dataset.plot(band=0)` renders a band inline (it returns a cleopatra ArrayGlyph). Dataset.read_file(tiles[0]).plot(band=0)

Out[3]:

<cleopatra.array_glyph.ArrayGlyph at 0x7ffa55e33cb0>

Mosaic — merge_rasters¶

Writes one raster covering the inputs' combined extent. method controls how overlaps resolve ('last', 'first', …).

In [4]:

mosaic = WORK / 'mosaic.tif'
merge_rasters(tiles, mosaic)
Dataset.read_file(mosaic).shape

mosaic = WORK / 'mosaic.tif' merge_rasters(tiles, mosaic) Dataset.read_file(mosaic).shape

Out[4]:

(1, 24, 29)

In [5]:

# The mosaic covers the inputs' combined extent — plot it to see the merged raster.
Dataset.read_file(mosaic).plot(band=0)

# The mosaic covers the inputs' combined extent — plot it to see the merged raster. Dataset.read_file(mosaic).plot(band=0)

Out[5]:

<cleopatra.array_glyph.ArrayGlyph at 0x7ffb08b48cd0>

Stack bands — stack_bands¶

Combine single-band rasters into one multi-band raster (order = input order).

In [6]:

stacked = WORK / 'stacked.tif'
stack_bands(tiles, path=stacked, band_names=['t0', 't1', 't2'])
out = Dataset.read_file(stacked)
out.band_count, out.band_names

stacked = WORK / 'stacked.tif' stack_bands(tiles, path=stacked, band_names=['t0', 't1', 't2']) out = Dataset.read_file(stacked) out.band_count, out.band_names

Out[6]:

(3, ['t0', 't1', 't2'])

In [7]:

# The stacked raster carries the three tiles as bands t0/t1/t2 — plot each band.
for b in range(out.band_count):
    out.plot(band=b)

# The stacked raster carries the three tiles as bands t0/t1/t2 — plot each band. for b in range(out.band_count): out.plot(band=b)

Notes¶

• merge_rasters takes no_data_value= and method= to control fill and overlap handling.
• stack_bands can align=True to snap mismatched grids before stacking.
• See also: Reproject / resample / align.