GEE 案例——如何计算sentinel-2中每一个单景影像的波段的DN值并绘制直方图

此星光明

发布于 2024-02-07 08:06:15

1260

发布于 2024-02-07 08:06:15

文章被收录于专栏：GEE数据专栏，GEE学习专栏，GEE错误集等专栏GEE数据专栏，GEE学习专栏，GEE错误集等专栏

原始问题

我正试图在 Google 地球引擎中为整个图像集合计算一个直方图。为了达到我想要的结果，我现在所做的是计算每个单独图像的直方图直方图1 并将它们相加，不知道是否正确。

简介

直方图基本上是一个配对值列表。因此，您可以用函数映射它，而无需 for/ 循环。以下代码片段包含了为整个图像集生成直方图的算法的重要部分。

函数

ee.Reducer.fixedHistogram(min, max, steps, cumulative)

Creates a reducer that will compute a histogram of the inputs using a fixed number of fixed width bins. Values outside of the [min, max) range are ignored. The output is a Nx2 array of bucket lower edges and counts (or cumulative counts) and is suitable for use per-pixel.

创建一个聚类器，使用固定数量、固定宽度的分隔来计算输入的直方图。超出 [min, max] 范围的值将被忽略。输出是一个 Nx2 数组，包含桶下边缘和计数（或累计计数），适合按像素使用。

Arguments:

min (Float):

The lower (inclusive) bound of the first bucket.

max (Float):

The upper (exclusive) bound of the last bucket.

steps (Integer):

The number of buckets to use.

cumulative (Boolean, default: false):

When true, generates a cumulative histogram.

Returns: Reducer

ui.Chart.image.histogram(image, region, scale, maxBuckets, minBucketWidth, maxRaw, maxPixels)

Generates a Chart from an image. Computes and plots histograms of the values of the bands in the specified region of the image.

X-axis: Histogram buckets (of band value).
Y-axis: Frequency (number of pixels with a band value in the bucket).

根据图像生成图表。计算并绘制图像指定区域内色带值的直方图。

X 轴直方图桶（带值）。

Y 轴频率（带值在桶中的像素数量）。

Returns a chart.

Arguments:

image (Image):

The image to generate a histogram from.

region (Feature|FeatureCollection|Geometry, optional):

The region to reduce. If omitted, uses the entire image.

scale (Number, optional):

The pixel scale used when applying the histogram reducer, in meters.

maxBuckets (Number, optional):

The maximum number of buckets to use when building a histogram; will be rounded up to a power of 2.

minBucketWidth (Number, optional):

The minimum histogram bucket width, or null to allow any power of 2.

maxRaw (Number, optional):

The number of values to accumulate before building the initial histogram.

maxPixels (Number, optional):

If specified, overrides the maximum number of pixels allowed in the histogram reduction. Defaults to 1e6.

Returns: ui.Chart

ui.Chart.array.values(array, axis, xLabels)

Generates a Chart from an array. Plots separate series for each 1-D vector along the given axis.

X-axis = Array index along axis, optionally labeled by xLabels.
Y-axis = Value.
Series = Vector, described by indices of the non-axis array axes.

从数组生成图表。沿着给定的坐标轴为每个一维向量绘制单独的序列。

X-axis = 沿轴的数组索引，可选择用 xLabels 标注。

Y 轴 = 数值。

系列 = 矢量，由非轴数组轴的索引描述。

Returns a chart.

Arguments:

array (Array|List<Object>):

Array to chart.

axis (Number):

The axis along which to generate the 1-D vector series.

xLabels (Array|List<Object>, optional):

Labels for ticks along the x-axis of the chart.

Returns: ui.Chart

代码：

// FUNCTIONS//
function normalize(image, minValue, maxValue){
    var image1 = image.subtract(minValue);
    var diff = maxValue.subtract(minValue);
    var image2 =  image1.divide(diff);
    return image2.copyProperties(image, ["system:time_start"]);
}

// CODE //

// Dataset selection
var s2 = ee.ImageCollection('COPERNICUS/S2').filterMetadata('MGRS_TILE', 'equals', '32UPU');
var start = ee.Date('2021-11-01');
var end = ee.Date('2022-01-01');
var s2 = s2.filterDate(start, end);
print("s2", s2);

// Normalizing image collection images with overall minimum and maximum - here only for Band 2
var band = 'B2';
var colIm = s2.select(band);
var minImage=colIm.min();
var maxImage=colIm.max();
var maxValue = ee.Number(maxImage.reduceRegion({
  reducer: ee.Reducer.max(),
  geometry: s2.geometry(),
  maxPixels:10e10,
  scale: 100
  }).get(band));
var minValue = ee.Number(minImage.reduceRegion({
  reducer : ee.Reducer.min(),
  geometry : s2.geometry(),
  maxPixels : 10e10,
  scale : 100
  }).get(band));
  
var s2_norm = colIm.map(function(img){return normalize(img, minValue, maxValue)});

print("s2_norm", s2_norm);
Map.centerObject(s2_norm);

// Histogram for single Image
var image = s2_norm.first();
var histogram_single = image.reduceRegion({
      reducer : ee.Reducer.fixedHistogram(0,1,256),
      geometry : s2_norm.geometry(),
      maxPixels : 10e9
      }).get(band);

print('Histogram for Single Image', histogram_single);

var values = ee.Array(histogram_single).toList().map(function (ele) {
  
  return ee.List(ele).get(0);
  
});

//print("values", values);

var counts = ee.Array(histogram_single).toList().map(function (ele) {
  
  return ee.List(ele).get(1);
  
});

//print("counts", counts);

// Create Histogram over all Images
var histo = s2_norm.reduce(ee.Reducer.fixedHistogram ({min:0, max: 1, steps: 256, cumulative: true}));
print('Output of Histogram for ImageCollection',histo.select('B2_histogram'));

Map.addLayer(s2_norm.geometry(), {}, 's2_norm');
//Map.addLayer(histo);

print("scale", image.projection().nominalScale());

// Define the chart and print it to the console.
var chart =
    ui.Chart.image.histogram({image: image, region: s2_norm.geometry(), scale: 35})
        .setOptions({
          title: 'Sentinel 2 normalized B2 band Histogram',
          hAxis: {
            title: 'Normalized values',
            titleTextStyle: {italic: false, bold: true},
            viewWindow: {min: 0, max: 1}
          },
          vAxis:
              {title: 'Count', titleTextStyle: {italic: false, bold: true}},
          colors: ['blue']
        });
print(chart);

var chart2 = ui.Chart.array.values({array: counts, axis: 0, xLabels: values})
                .setOptions({
                  title: 'Sentinel 2 normalized B2 band Histogram',
                  colors: ['red'],
                  hAxis: {
                    'title': 'Normalized values',
                    titleTextStyle: {italic: false, bold: true}
                  },
                  vAxis: 
                    {title: 'Count', titleTextStyle: {italic: false, bold: true}},
                  
                }).setChartType('ColumnChart');

print(chart2);

var s2_norm_lst = s2_norm.toList(s2_norm.size());

var allValues = s2_norm_lst.map(function (ele) {
  
  var histogram = ee.Image(ele).reduceRegion({
        reducer: ee.Reducer.fixedHistogram(0,1,255),
        geometry: s2_norm.geometry(),
        maxPixels: 10e9}
      ).get(band);
  
  return ee.List(ee.Array(histogram).toList());
  
});

allValues = ee.List(allValues).flatten();

var len = allValues.size();

var list = ee.List.sequence(0, len.subtract(1), 2);

var tot_list = list.map(function(ele){

  var start = ee.Number(ele).int(); 
  var end = ee.Number(ele).add(2).int(); 

  var new_list = ee.List([]);
  var element = ee.List(allValues.slice(start, end));

  new_list = new_list.add(element);

  return new_list.get(0);

});

//print(tot_list);

var allValues = tot_list.map(function (ele) {
  
  return ee.List(ele).get(0);
  
});

var allCounts = tot_list.map(function (ele) {
  
  return ee.List(ele).get(1);
  
});

var chart3 = ui.Chart.array.values({array: allCounts, axis: 0, xLabels: allValues})
                .setOptions({
                  title: 'Sentinel 2 normalized B2 band Histogram',
                  colors: ['red'],
                  hAxis: {
                    'title': 'Normalized values',
                    titleTextStyle: {italic: false, bold: true}
                  },
                  vAxis: 
                    {title: 'Count', titleTextStyle: {italic: false, bold: true}},
                  
                }).setChartType('ColumnChart');

print(chart3);

第一张图表（蓝色）是使用集合中第一张图像的 ui.Chart.image.histogram 获得的（您的 histo 图像对于获得整个集合的直方图没有用处，也无法添加到地图画布中）。只能在比例尺大于 35 米时使用（但是，B2 波段的比例尺为 10 米）。第二和第三个图表（红色）分别显示第一幅图像和整个图像集。

代码2

// FUNCTIONS//
function normalize(image, minValue, maxValue){
    var image1 = image.subtract(minValue);
    var diff = maxValue.subtract(minValue);
    var image2 =  image1.divide(diff);
    return image2.copyProperties(image, ["system:time_start"]);
}

// CODE //

// Dataset selection
var s2 = ee.ImageCollection('COPERNICUS/S2').filterMetadata('MGRS_TILE', 'equals', '32UPU');
var start = ee.Date('2021-11-01');
var end = ee.Date('2022-01-01');
var s2 = s2.filterDate(start, end);
print(s2);

// Normalizing image collection images with overall minimum and maximum - here only for Band 2
var band = 'B2';
var colIm = s2.select(band);
var minImage=colIm.min();
var maxImage=colIm.max();
var maxValue = ee.Number(maxImage.reduceRegion({
  reducer: ee.Reducer.max(),
  geometry: s2.geometry(),
  maxPixels:10e10,
  scale: 100
  }).get(band));
var minValue = ee.Number(minImage.reduceRegion({
  reducer : ee.Reducer.min(),
  geometry : s2.geometry(),
  maxPixels : 10e10,
  scale : 100
  }).get(band));
  
var s2_norm = colIm.map(function(img){return normalize(img, minValue, maxValue)});

// Histogram for single Image
var image = s2_norm.first();
var histogram_single = image.reduceRegion({
      reducer : ee.Reducer.fixedHistogram(0,1,255),
      geometry : s2_norm.geometry(),
      maxPixels : 10e9
      }).get(band);
print('Histogram for Single Image', histogram_single);

// Create Histogram over all Images
var s2_norm_lst = s2_norm.toList(s2_norm.size());

var allValues = s2_norm_lst.map(function (ele) {
  var histogram = ee.Image(ele).reduceRegion({
        reducer: ee.Reducer.fixedHistogram(0,1,255),
        geometry: s2_norm.geometry(),
        maxPixels: 10e9}
      ).get(band);
  return ee.List(ee.Array(histogram).toList());
});


allValues = ee.List(allValues).flatten();

var histo = ee.List([]);
for (var i=0;i<255;i++){
  var slice = ee.List(allValues.slice(i*2+1, ee.Number(allValues.size()),510));
  var element = ee.Number(slice.reduce(ee.Reducer.sum()));
  var bin = ee.List(allValues.slice(i*2, ee.Number(allValues.size()),510));
  var element2 = ee.Number(bin.reduce(ee.Reducer.mean()));
  var histo = histo.add([element2,element]);
}

print('Output of Histogram for ImageCollection', histo);

var allValues = histo.map(function (ele) {
  
  return ee.List(ele).get(0);
  
});

var allCounts = histo.map(function (ele) {
  
  return ee.List(ele).get(1);
  
});

var chart3 = ui.Chart.array.values({array: allCounts, axis: 0, xLabels: allValues})
                .setOptions({
                  title: 'Sentinel 2 normalized B2 band Histogram',
                  colors: ['red'],
                  hAxis: {
                    'title': 'Normalized values',
                    titleTextStyle: {italic: false, bold: true}
                  },
                  vAxis: 
                    {title: 'Count', titleTextStyle: {italic: false, bold: true}},
                  
                }).setChartType('ColumnChart');

print(chart3);