ee.Classifier.smileKNN
Restez organisé à l'aide des collections
Enregistrez et classez les contenus selon vos préférences.
Crée un classificateur k-NN vide.
L'algorithme des k plus proches voisins (k-NN) est une méthode de classification des objets par vote majoritaire de leurs voisins. L'objet est attribué à la classe la plus fréquente parmi ses k plus proches voisins (k est un entier positif, généralement petit et impair).
Utilisation | Renvoie |
---|
ee.Classifier.smileKNN(k, searchMethod, metric) | Classificateur |
Argument | Type | Détails |
---|
k | Entier, valeur par défaut : 1 | Nombre de voisins pour la classification. |
searchMethod | Chaîne, valeur par défaut : "AUTO" | Méthode de recherche. Les valeurs valides sont [AUTO, LINEAR_SEARCH, KD_TREE, COVER_TREE].
AUTO choisira entre KD_TREE et COVER_TREE en fonction du nombre de dimensions. Les résultats peuvent varier entre les différentes méthodes de recherche pour les égalités de distance et les valeurs de probabilité. Étant donné que les performances et les résultats peuvent varier, consultez la documentation de SMILE et d'autres ressources. |
metric | Chaîne, valeur par défaut : "EUCLIDEAN" | Métrique de distance à utiliser. REMARQUE : KD_TREE (et AUTO pour les dimensions faibles) n'utilisera pas la métrique sélectionnée. Les options sont les suivantes :
"EUCLIDEAN" : distance euclidienne.
'MAHALANOBIS' : distance de Mahalanobis.
"MANHATTAN" : distance de Manhattan.
'BRAYCURTIS' : distance de Bray-Curtis. |
Exemples
Éditeur de code (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']});
Configuration de Python
Consultez la page
Environnement Python pour en savoir plus sur l'API Python et sur l'utilisation de geemap
pour le développement interactif.
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
Sauf indication contraire, le contenu de cette page est régi par une licence Creative Commons Attribution 4.0, et les échantillons de code sont régis par une licence Apache 2.0. Pour en savoir plus, consultez les Règles du site Google Developers. Java est une marque déposée d'Oracle et/ou de ses sociétés affiliées.
Dernière mise à jour le 2025/07/26 (UTC).
[null,null,["Dernière mise à jour le 2025/07/26 (UTC)."],[[["\u003cp\u003eCreates a k-Nearest Neighbors (k-NN) classifier using the SMILE machine learning library within Google Earth Engine.\u003c/p\u003e\n"],["\u003cp\u003eThe classifier is trained using labeled data and can be applied to classify images based on the proximity of pixel values to known classes.\u003c/p\u003e\n"],["\u003cp\u003eUsers can customize the number of neighbors (k), search method, and distance metric for the k-NN algorithm.\u003c/p\u003e\n"],["\u003cp\u003eIncludes JavaScript and Python examples demonstrating classifier training and image classification using Landsat 8 data.\u003c/p\u003e\n"]]],[],null,["# ee.Classifier.smileKNN\n\nCreates an empty k-NN classifier.\n\n\u003cbr /\u003e\n\nThe 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).\n\n| Usage | Returns |\n|-----------------------------------------------------------------|------------|\n| `ee.Classifier.smileKNN(`*k* `, `*searchMethod* `, `*metric*`)` | Classifier |\n\n| Argument | Type | Details |\n|----------------|------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| `k` | Integer, default: 1 | The number of neighbors for classification. |\n| `searchMethod` | String, 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. |\n| `metric` | String, 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. |\n\nExamples\n--------\n\n### Code Editor (JavaScript)\n\n```javascript\n// Cloud masking for Landsat 8.\nfunction maskL8sr(image) {\n var qaMask = image.select('QA_PIXEL').bitwiseAnd(parseInt('11111', 2)).eq(0);\n var saturationMask = image.select('QA_RADSAT').eq(0);\n\n // Apply the scaling factors to the appropriate bands.\n var opticalBands = image.select('SR_B.').multiply(0.0000275).add(-0.2);\n var thermalBands = image.select('ST_B.*').multiply(0.00341802).add(149.0);\n\n // Replace the original bands with the scaled ones and apply the masks.\n return image.addBands(opticalBands, null, true)\n .addBands(thermalBands, null, true)\n .updateMask(qaMask)\n .updateMask(saturationMask);\n}\n\n// Map the function over one year of data.\nvar collection = ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')\n .filterDate('2020-01-01', '2021-01-01')\n .map(maskL8sr);\n\n// Make a median composite.\nvar composite = collection.median();\n\n// Demonstration labels.\nvar labels = ee.FeatureCollection('projects/google/demo_landcover_labels')\n\n// Use these bands for classification.\nvar bands = ['SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7'];\n// The name of the property on the points storing the class label.\nvar classProperty = 'landcover';\n\n// Sample the composite to generate training data. Note that the\n// class label is stored in the 'landcover' property.\nvar training = composite.select(bands).sampleRegions(\n {collection: labels, properties: [classProperty], scale: 30});\n\n// Train a kNN classifier.\nvar classifier = ee.Classifier.smileKNN(5).train({\n features: training,\n classProperty: classProperty,\n});\n\n// Classify the composite.\nvar classified = composite.classify(classifier);\nMap.setCenter(-122.184, 37.796, 12);\nMap.addLayer(classified, {min: 0, max: 2, palette: ['red', 'green', 'blue']});\n```\nPython setup\n\nSee the [Python Environment](/earth-engine/guides/python_install) page for information on the Python API and using\n`geemap` for interactive development. \n\n```python\nimport ee\nimport geemap.core as geemap\n```\n\n### Colab (Python)\n\n```python\n# Cloud masking for Landsat 8.\ndef mask_l8_sr(image):\n qa_mask = image.select('QA_PIXEL').bitwiseAnd(int('11111', 2)).eq(0)\n saturation_mask = image.select('QA_RADSAT').eq(0)\n\n # Apply the scaling factors to the appropriate bands.\n optical_bands = image.select('SR_B.').multiply(0.0000275).add(-0.2)\n thermal_bands = image.select('ST_B.*').multiply(0.00341802).add(149.0)\n\n # Replace the original bands with the scaled ones and apply the masks.\n return (\n image.addBands(optical_bands, None, True)\n .addBands(thermal_bands, None, True)\n .updateMask(qa_mask)\n .updateMask(saturation_mask)\n )\n\n\n# Map the function over one year of data.\ncollection = (\n ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')\n .filterDate('2020-01-01', '2021-01-01')\n .map(mask_l8_sr)\n)\n\n# Make a median composite.\ncomposite = collection.median()\n\n# Demonstration labels.\nlabels = ee.FeatureCollection('projects/google/demo_landcover_labels')\n\n# Use these bands for classification.\nbands = ['SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7']\n# The name of the property on the points storing the class label.\nclass_property = 'landcover'\n\n# Sample the composite to generate training data. Note that the\n# class label is stored in the 'landcover' property.\ntraining = composite.select(bands).sampleRegions(\n collection=labels, properties=[class_property], scale=30\n)\n\n# Train a kNN classifier.\nclassifier = ee.Classifier.smileKNN(5).train(\n features=training, classProperty=class_property\n)\n\n# Classify the composite.\nclassified = composite.classify(classifier)\n\nm = geemap.Map()\nm.set_center(-122.184, 37.796, 12)\nm.add_layer(\n classified, {'min': 0, 'max': 2, 'palette': ['red', 'green', 'blue']}\n)\nm\n```"]]