ee.Classifier.smileKNN

Creates an empty k-NN classifier.

The k-nearest neighbor algorithm (k-NN) is a method for classifying objects by a majority vote of its neighbors, with the object being assigned to the class most common amongst its k nearest neighbors (k is a positive integer, typically small, typically odd).

UsageReturns
ee.Classifier.smileKNN(k, searchMethod, metric)Classifier
ArgumentTypeDetails
kInteger, default: 1The number of neighbors for classification.
searchMethodString, default: "AUTO"Search method. The following are valid [AUTO, LINEAR_SEARCH, KD_TREE, COVER_TREE]. AUTO will choose between KD_TREE and COVER_TREE depending on the dimension count. Results may vary between the different search methods for distance ties and probability values. Since performance and results may vary consult with SMILE's documentation and other literature.
metricString, default: "EUCLIDEAN"The distance metric to use. NOTE: KD_TREE (and AUTO for low dimensions) will not use the metric selected. Options are:   'EUCLIDEAN' - Euclidean distance.   'MAHALANOBIS' - Mahalanobis distance.   'MANHATTAN' - Manhattan distance.   'BRAYCURTIS' - Bray-Curtis distance.

Examples

Code Editor (JavaScript)

// Cloud masking for Landsat 8.
function maskL8sr(image) {
  var qaMask = image.select('QA_PIXEL').bitwiseAnd(parseInt('11111', 2)).eq(0);
  var saturationMask = image.select('QA_RADSAT').eq(0);

  // Apply the scaling factors to the appropriate bands.
  var opticalBands = image.select('SR_B.').multiply(0.0000275).add(-0.2);
  var thermalBands = image.select('ST_B.*').multiply(0.00341802).add(149.0);

  // Replace the original bands with the scaled ones and apply the masks.
  return image.addBands(opticalBands, null, true)
      .addBands(thermalBands, null, true)
      .updateMask(qaMask)
      .updateMask(saturationMask);
}

// Map the function over one year of data.
var collection = ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')
                     .filterDate('2020-01-01', '2021-01-01')
                     .map(maskL8sr);

// Make a median composite.
var composite = collection.median();

// Demonstration labels.
var labels = ee.FeatureCollection('projects/google/demo_landcover_labels')

// Use these bands for classification.
var bands = ['SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7'];
// The name of the property on the points storing the class label.
var classProperty = 'landcover';

// Sample the composite to generate training data.  Note that the
// class label is stored in the 'landcover' property.
var training = composite.select(bands).sampleRegions(
    {collection: labels, properties: [classProperty], scale: 30});

// Train a kNN classifier.
var classifier = ee.Classifier.smileKNN(5).train({
  features: training,
  classProperty: classProperty,
});

// Classify the composite.
var classified = composite.classify(classifier);
Map.setCenter(-122.184, 37.796, 12);
Map.addLayer(classified, {min: 0, max: 2, palette: ['red', 'green', 'blue']});

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)

# Cloud masking for Landsat 8.
def mask_l8_sr(image):
  qa_mask = image.select('QA_PIXEL').bitwiseAnd(int('11111', 2)).eq(0)
  saturation_mask = image.select('QA_RADSAT').eq(0)

  # Apply the scaling factors to the appropriate bands.
  optical_bands = image.select('SR_B.').multiply(0.0000275).add(-0.2)
  thermal_bands = image.select('ST_B.*').multiply(0.00341802).add(149.0)

  # Replace the original bands with the scaled ones and apply the masks.
  return (
      image.addBands(optical_bands, None, True)
      .addBands(thermal_bands, None, True)
      .updateMask(qa_mask)
      .updateMask(saturation_mask)
  )


# Map the function over one year of data.
collection = (
    ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')
    .filterDate('2020-01-01', '2021-01-01')
    .map(mask_l8_sr)
)

# Make a median composite.
composite = collection.median()

# Demonstration labels.
labels = ee.FeatureCollection('projects/google/demo_landcover_labels')

# Use these bands for classification.
bands = ['SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7']
# The name of the property on the points storing the class label.
class_property = 'landcover'

# Sample the composite to generate training data.  Note that the
# class label is stored in the 'landcover' property.
training = composite.select(bands).sampleRegions(
    collection=labels, properties=[class_property], scale=30
)

# Train a kNN classifier.
classifier = ee.Classifier.smileKNN(5).train(
    features=training, classProperty=class_property
)

# Classify the composite.
classified = composite.classify(classifier)

m = geemap.Map()
m.set_center(-122.184, 37.796, 12)
m.add_layer(
    classified, {'min': 0, 'max': 2, 'palette': ['red', 'green', 'blue']}
)
m