Overview
The Geometric Median function calculates the geometric median across pixels in a time series of multiband imagery.
Notes
This function reduces noise and outliers in time series imagery by calculating a geometric median pixel for each pixel array across the image stack. The algorithm maintains the spectral relationships between bands in the pixel spectra, so the output can be used in analysis, such as vegetation indices.
Clouds and shadows should be masked using the dataset QA band before applying this function.
If the input image is floating point, such as surface reflectance with values between 0 and 1, an epsilon value of 0.001 should produce quality results.
Parameters
Parameter name  Description 

Rasters  The input multiband rasters. 
Cell Size Type  Choose which cell size to use in the output raster. If all the input cell sizes are the same, all the options will yield the same results.

Extent Type  Choose which extent should be used in the output raster:

Maximum number of iterations  Specifies the maximum number of iterations to complete. The computation ends when this value is reached, regardless of the Epsilon setting. 
Epsilon  Specifies the convergence value between two consecutive iterations. When epsilon is less than or equal to the specified value, the iteration stops, and the result of the last iteration is used. 