homonim.RasterFuse.process

RasterFuse.process(corr_filename: Path | str, model: Model = Model.gain_blk_offset, kernel_shape: Tuple[int, int] = (5, 5), param_filename: Path | str | None = None, build_ovw: bool = True, overwrite: bool = False, model_config: Dict | None = None, out_profile: Dict | None = None, block_config: Dict | None = None)

Correct source image to surface reflectance by fusion with a reference.

To improve speed and reduce memory usage, images are divided into blocks for concurrent processing.

Parameters:

  • corr_filename (str, Path) – Path to the corrected file to create.

  • model (homonim.enums.Model, optional) – Correction model to use. See Model for options.

  • kernel_shape (tuple of int, optional) – (height, width) of the kernel in pixels of the proc_crs image (the lowest resolution image, if proc_crs = auto). Larger kernels are less susceptible to over-fitting on noisy data, but provide lower resolution correction.

  • param_filename (str, Path, optional) – Path to an optional parameter file to write with correction parameters and R2 values. By default, no parameter file is written.

  • build_ovw (bool, optional) – Build overviews for the output image(s).

  • overwrite (bool, optional) – Overwrite the output image(s) if they exist.

  • model_config (dict, optional) – Configuration dictionary for the correction model. See create_model_config() for keys and default values.

  • out_profile (dict, optional) – Configuration dictionary for the output image(s). See create_out_profile() for keys and default values.

  • block_config (dict, optional) – Configuration dictionary for block processing. See create_block_config() for keys and default values.