ee.FeatureCollection.cluster

对集合中的每个特征进行聚类，并为每个特征添加一个新列，其中包含已分配的聚类编号。

用法返回
FeatureCollection.cluster(clusterer, outputName)FeatureCollection
参数类型详细信息
此：featuresFeatureCollection要聚类的特征集合。每个特征都必须包含聚类器架构中的所有属性。
clusterer聚类器要使用的聚类器。
outputName字符串，默认值：“cluster”要添加的输出属性的名称。

示例

代码编辑器 (JavaScript)

// Import a Sentinel-2 surface reflectance image.
var image = ee.Image('COPERNICUS/S2_SR/20210109T185751_20210109T185931_T10SEG');

// Get the image geometry to define the geographical bounds of a point sample.
var imageBounds = image.geometry();

// Sample the image at a set of random points; a feature collection is returned.
var pointSampleFc = image.sample(
    {region: imageBounds, scale: 20, numPixels: 1000, geometries: true});

// Instantiate a k-means clusterer and train it.
var clusterer = ee.Clusterer.wekaKMeans(5).train(pointSampleFc);

// Cluster the input using the trained clusterer; optionally specify the name
// of the output cluster ID property.
var clusteredFc = pointSampleFc.cluster(clusterer, 'spectral_cluster');

print('Note added "spectral_cluster" property for an example feature',
      clusteredFc.first().toDictionary());

// Visualize the clusters by applying a unique color to each cluster ID.
var palette = ee.List(['8dd3c7', 'ffffb3', 'bebada', 'fb8072', '80b1d3']);
var clusterVis = clusteredFc.map(function(feature) {
  return feature.set('style', {
    color: palette.get(feature.get('spectral_cluster')),
  });
}).style({styleProperty: 'style'});

// Display the points colored by cluster ID with the S2 image.
Map.setCenter(-122.35, 37.47, 9);
Map.addLayer(image, {bands: ['B4', 'B3', 'B2'], min: 0, max: 1500}, 'S2 image');
Map.addLayer(clusterVis, null, 'Clusters');

Python 设置

如需了解 Python API 和如何使用 geemap 进行交互式开发，请参阅 Python 环境页面。

import ee
import geemap.core as geemap

Colab (Python)

# Import a Sentinel-2 surface reflectance image.
image = ee.Image('COPERNICUS/S2_SR/20210109T185751_20210109T185931_T10SEG')

# Get the image geometry to define the geographical bounds of a point sample.
image_bounds = image.geometry()

# Sample the image at a set of random points a feature collection is returned.
point_sample_fc = image.sample(
    region=image_bounds, scale=20, numPixels=1000, geometries=True
)

# Instantiate a k-means clusterer and train it.
clusterer = ee.Clusterer.wekaKMeans(5).train(point_sample_fc)

# Cluster the input using the trained clusterer optionally specify the name
# of the output cluster ID property.
clustered_fc = point_sample_fc.cluster(clusterer, 'spectral_cluster')

display(
    'Note added "spectral_cluster" property for an example feature',
    clustered_fc.first().toDictionary(),
)

# Visualize the clusters by applying a unique color to each cluster ID.
palette = ee.List(['8dd3c7', 'ffffb3', 'bebada', 'fb8072', '80b1d3'])
cluster_vis = clustered_fc.map(
    lambda feature: feature.set(
        'style', {'color': palette.get(feature.get('spectral_cluster'))}
    )
).style(styleProperty='style')

# Display the points colored by cluster ID with the S2 image.
m = geemap.Map()
m.set_center(-122.35, 37.47, 9)
m.add_layer(
    image, {'bands': ['B4', 'B3', 'B2'], 'min': 0, 'max': 1500}, 'S2 image'
)
m.add_layer(cluster_vis, None, 'Clusters')
m