ee.ImageCollection.median

Reduces an image collection by calculating the median of all values at each pixel across the stack of all matching bands. Bands are matched by name.

Usage	Returns
ImageCollection.median()	Image

Argument	Type	Details
this: collection	ImageCollection	The image collection to reduce.

Examples

Code Editor (JavaScript)

// Sentinel-2 image collection for July 2021 intersecting a point of interest.
// Reflectance, cloud probability, and scene classification bands are selected.
var col = ee.ImageCollection('COPERNICUS/S2_SR')
  .filterDate('2021-07-01', '2021-08-01')
  .filterBounds(ee.Geometry.Point(-122.373, 37.448))
  .select('B.*|MSK_CLDPRB|SCL');

// Visualization parameters for reflectance RGB.
var visRefl = {
  bands: ['B11', 'B8', 'B3'],
  min: 0,
  max: 4000
};
Map.setCenter(-122.373, 37.448, 9);
Map.addLayer(col, visRefl, 'Collection reference', false);

// Reduce the collection to a single image using a variety of methods.
var mean = col.mean();
Map.addLayer(mean, visRefl, 'Mean (B11, B8, B3)');

var median = col.median();
Map.addLayer(median, visRefl, 'Median (B11, B8, B3)');

var min = col.min();
Map.addLayer(min, visRefl, 'Min (B11, B8, B3)');

var max = col.max();
Map.addLayer(max, visRefl, 'Max (B11, B8, B3)');

var sum = col.sum();
Map.addLayer(sum,
  {bands: ['MSK_CLDPRB'], min: 0, max: 500}, 'Sum (MSK_CLDPRB)');

var product = col.product();
Map.addLayer(product,
  {bands: ['MSK_CLDPRB'], min: 0, max: 1e10}, 'Product (MSK_CLDPRB)');

// ee.ImageCollection.mode returns the most common value. If multiple mode
// values occur, the minimum mode value is returned.
var mode = col.mode();
Map.addLayer(mode, {bands: ['SCL'], min: 1, max: 11}, 'Mode (pixel class)');

// ee.ImageCollection.count returns the frequency of valid observations. Here,
// image pixels are masked based on cloud probability to add valid observation
// variability to the collection. Note that pixels with no valid observations
// are masked out of the returned image.
var notCloudCol = col.map(function(img) {
  return img.updateMask(img.select('MSK_CLDPRB').lte(10));
});
var count = notCloudCol.count();
Map.addLayer(count, {min: 1, max: 5}, 'Count (not cloud observations)');

// ee.ImageCollection.mosaic composites images according to their position in
// the collection (priority is last to first) and pixel mask status, where
// invalid (mask value 0) pixels are filled by preceding valid (mask value >0)
// pixels.
var mosaic = notCloudCol.mosaic();
Map.addLayer(mosaic, visRefl, 'Mosaic (B11, B8, B3)');

Python setup

See the Python Environment page for information on the Python API and using geemap for interactive development.

import ee
import geemap.core as geemap

Colab (Python)

# Sentinel-2 image collection for July 2021 intersecting a point of interest.
# Reflectance, cloud probability, and scene classification bands are selected.
col = (
    ee.ImageCollection('COPERNICUS/S2_SR')
    .filterDate('2021-07-01', '2021-08-01')
    .filterBounds(ee.Geometry.Point(-122.373, 37.448))
    .select('B.*|MSK_CLDPRB|SCL')
)

# Visualization parameters for reflectance RGB.
vis_refl = {'bands': ['B11', 'B8', 'B3'], 'min': 0, 'max': 4000}
m = geemap.Map()
m.set_center(-122.373, 37.448, 9)
m.add_layer(col, vis_refl, 'Collection reference', False)

# Reduce the collection to a single image using a variety of methods.
mean = col.mean()
m.add_layer(mean, vis_refl, 'Mean (B11, B8, B3)')

median = col.median()
m.add_layer(median, vis_refl, 'Median (B11, B8, B3)')

min = col.min()
m.add_layer(min, vis_refl, 'Min (B11, B8, B3)')

max = col.max()
m.add_layer(max, vis_refl, 'Max (B11, B8, B3)')

sum = col.sum()
m.add_layer(
    sum, {'bands': ['MSK_CLDPRB'], 'min': 0, 'max': 500}, 'Sum (MSK_CLDPRB)'
)

product = col.product()
m.add_layer(
    product,
    {'bands': ['MSK_CLDPRB'], 'min': 0, 'max': 1e10},
    'Product (MSK_CLDPRB)',
)

# ee.ImageCollection.mode returns the most common value. If multiple mode
# values occur, the minimum mode value is returned.
mode = col.mode()
m.add_layer(
    mode, {'bands': ['SCL'], 'min': 1, 'max': 11}, 'Mode (pixel class)'
)

# ee.ImageCollection.count returns the frequency of valid observations. Here,
# image pixels are masked based on cloud probability to add valid observation
# variability to the collection. Note that pixels with no valid observations
# are masked out of the returned image.
not_cloud_col = col.map(
    lambda img: img.updateMask(img.select('MSK_CLDPRB').lte(10))
)
count = not_cloud_col.count()
m.add_layer(count, {'min': 1, 'max': 5}, 'Count (not cloud observations)')

# ee.ImageCollection.mosaic composites images according to their position in
# the collection (priority is last to first) and pixel mask status, where
# invalid (mask value 0) pixels are filled by preceding valid (mask value >0)
# pixels.
mosaic = not_cloud_col.mosaic()
m.add_layer(mosaic, vis_refl, 'Mosaic (B11, B8, B3)')
m