PyramidingPolicy

  • The pyramiding policy determines how reduced-resolution versions of image bands are created, impacting the representation of pixel values at lower resolutions.

  • By default, the MEAN policy is used, calculating the output pixel as the average of input pixels, weighted by their masks, ideal for continuous data like temperature.

  • Alternative policies like SAMPLE, MIN, MAX, MODE, and MEDIAN offer different methods for pixel value determination, catering to various data types such as QA bands, categorical data, and more.

  • SAMPLE policy directly uses the upper-left input pixel for the output, suitable for non-continuous fields like date of acquisition or QA bitmask bands.

  • Users can choose the most appropriate policy depending on the nature of the data and desired outcome for image reduction.

The algorithm to be used when creating reduced-resolution versions of an image band. By default, the mask is computed as the MEAN of the input masks, reflecting the total weight of the input data over which the requested output statistic was calculated. In the case of SAMPLE, the sampled pixel's mask is used directly.

Enums
PYRAMIDING_POLICY_UNSPECIFIED Not specified.
MEAN Output pixel is the mask-weighted mean of the input pixels. Should be used for direct observation results and continuous quantities like temperature.
SAMPLE Output pixel is identical to the upper-left input pixel. Should be used for QA/bitmask bands or other non-continuous fields like date of pixel acquisition.
MIN Output pixel is the minimum value of the input pixels.
MAX Output pixel is the maximum value of the input pixels.
MODE Output pixel is the mask-weighted mode of the input pixels. Should be used for landcover bands.
MEDIAN Output pixel is the mask-weighted median of the input pixels. In case of a tie, the mean of the middle two values is used.