Resize a geospatial image

Resize a GEO, or DSM object to a desired pixel size. Wraps plantcv.plantcv.transform.resize and updates the affine transform on GEO and DSM objects to preserve geospatial coordinates after resizing.

plantcv.geospatial.resize(img, size, interpolation="auto")

returns Resized image of the same class as the input

  • Parameters:

    • img - Image, GEO, or DSM object, typically read in with gcv.read_geotif
    • size - Output image size in pixels as a tuple (width, height)
    • interpolation - Interpolation method (default: "auto"):
      • "auto" = automatically select "bicubic" when enlarging, "area" when reducing
      • "area" = resampling using pixel area relation (recommended for shrinking)
      • "bicubic" = bicubic interpolation over a 4×4 pixel neighborhood
      • "bilinear" = bilinear interpolation
      • "lanczos" = Lanczos interpolation over an 8×8 pixel neighborhood
      • "nearest" = nearest-neighbor interpolation (recommended for masks)
      • None = no interpolation; crop or zero-pad to reach the target size

  • Context:

    • When resizing a GEO or DSM object, the affine transform is scaled so that geospatial coordinates remain valid at the new resolution. The top-left corner of the image is preserved; only the pixel size components of the transform are updated.

  • Note:

    • The default interpolation "auto" works well for color GEO images. However, we recommend using "nearest" for DSMs.
import plantcv.geospatial as gcv

# Read a multispectral geotif
ortho = gcv.read_geotif(filename="./data/example_img.tif", bands="b,g,r,RE,NIR")

# Resize to 2000 x 2000 pixels using the default auto interpolation
resized = gcv.resize(img=ortho, size=(2000, 2000))

# Read a DSM
dsm = gcv.read_geotif(filename="./data/example_dsm.tif", bands="gray")

# Resize to 2000 x 2000 pixels using nearest-neighbor interpolation
resized = gcv.resize(img=dsm, size=(2000, 2000), interpolation="nearest")

Before Screenshot

After Screenshot

Source Code: Here