FROM_ GLC introduction and data download tutorial

1. Data is introduced

1.1 Data brief

Gong Peng's team is based on Google Earth Engine platform , Use globally 13 Ten thousand sample points , The random forest classification method is used to classify Sentinel-2 image , the 2017 Year of 10 M resolution global land cover mapping . The data set is divided into 10 class .

Category	Name	Code
Arable land,	Cropland	10
Woodland	Forest	20
The grass	Grassland	30
shrub	Shrubland	40
wetland	Wetland	50
Water	Water	60
Permafrost	Tundra	70
Impervious surface	Impervious surface	80
Naked land	Bareland	90
snow / ice	Snow/Ice	100

1.2 Data accuracy

The overall accuracy of this data set is 72.76%, User accuracy for each category 、 The producer accuracy is shown in the following table ：

2. Data download

Here are two data download methods

2.1 Download method on the official website

2.1.1 Download url

http://data.ess.tsinghua.edu.cn/fromglc2017v1.html, Open to directly enter the download interface , No registration and review required .

2.1.2 Image determination

The website provides data download according to the longitude and latitude of the image , The meanings of image names are as follows ：

Longitude and latitude are the coordinates of the bottom left corner of the image ：

The coordinates of each land cover image on the official website are increasing by an even number , The highest latitude is the North-South latitude 84°：

latitude ：-84,-82…-4,-2,0,2,4…82,84

longitude ：-180,-178,…,-4,-2,0,2,4,…178

Suppose we download the image of the center of Beijing （ North latitude 39°56′, East longitude 116°20′）, What you need to download is the north latitude 38°, East longitude 116° The image of ：

Download the video and remember two things ： The image name represents the latitude and longitude of the lower left corner ; Longitude and latitude increase by an even number .

2.2 Remote sensing cloud computing download method

Here is the second image download method ,PIE-Engine.PIE-Engine There are FROM_GLC Of 10 Meter resolution data (2017 year ).

There's this data , You can use it directly PIE-Engine Export... Tailored to administrative boundaries 10 Meter resolution land cover data .

2.2.1 The study area is ready to

stay PIE Upload the data of the research area , What I uploaded is Sichuan Province _ Ziyang City _ Administrative district data of Lezhi county .

After the data is uploaded , And GEE Different , The study area cannot directly call , Need to change to geometry type .

// Study area loading  LZ（ Sichuan Province _ Ziyang City _ Lezhi county ）
LZ= LZ.first().geometry(); 
Map.centerObject(LZ,9);    
Map.addLayer(LZ, {
    color: 'FF0000', fillColor: '00000000', width: 1}, "LZ")

2.2.2 Data filtering

First load the image set , Screening time , Select band , Inlay and cutting .

//FROM_GLC10  Apprentice overlay data download  
// Load apprentice overlay data   Screening 、 set 、 tailoring  
var LZ_landcover_2017=pie.ImageCollection("THU/FROM_GLC10_2017")
              .filterBounds(LZ)
              .filterDate("2017-1-01", "2017-12-31")
              .select(["B1"])
              .first()
              .clip(LZ);

// Show apprentice override data 
Map.addLayer(LZ_landcover_2017,visParam,"LZ_landcover_2017")

ad locum , There's a place to pay attention to :

PIE And GEE The cutting is different , You need to specify a band to crop .

2.2.3 Export data

It's almost the same as GEE almost , Select the download area 、 Download path, etc .

// download 
Export.image({
    
    image:LZ_landcover_2017,
    description: "LZ_landcover_2017",
    assetId: "LZ_landcover_2017",
    region:LZ,
    scale:10,
    maxPixels:1e13
});

There are two things to note :

ppqq If you don't know the size of the download area ,maxPixels Remember to make it bigger ;

ppqq PIE After the data is processed , stay PIE In resources , Click the file to download ;

2.2.4 Data usage

After the data is downloaded , Load into gis In the software , You can see the study area 2017 Year of 30 M land cover data .

10 M land cover data set , Than 30 Rice is much more delicate :

2.2.5 Code link

The code linking function provides two ways .
One is External links Method is used for non login users :
https://engine.piesat.cn/engine-share/shareCode.html?id=95120018f4ae488ea92c1f8b35b4c3ee
One is internal link The way , Open it directly in the editor for direct use by login users :
https://engine.piesat.cn/engine/home?sourceId=95120018f4ae488ea92c1f8b35b4c3ee

2.2.6 Complete code

// Study area loading   Sichuan Province _ Ziyang City _ Lezhi county 
LZ= LZ.first().geometry(); 
Map.centerObject(LZ,9);    
Map.addLayer(LZ, {
    color: 'FF0000', fillColor: '00000000', width: 1}, "LZ")

// Apprentice override data style 
var visParam = {
    
    min: 10,
    max: 100,
    palette: '#43B87C,#24DB99,#7EB451,#2E79BC,#2838B8,#8B2CC0,#EEE912,#BC1FA1,#17214F,#B81A74,#B5CF52,#932626,#2B328B,#AA5C5C,#2561E9,#874949,#4ECF61,#AE5151'
};

//FROM_GLC10  Apprentice overlay data download  
// Load apprentice overlay data   Screening 、 set 、 tailoring  
var LZ_landcover_2017=pie.ImageCollection("THU/FROM_GLC10_2017")
              .filterBounds(LZ)
              .filterDate("2017-1-01", "2017-12-31")
              .select(["B1"])
              .first()
              .clip(LZ);

// Show apprentice override data 
Map.addLayer(LZ_landcover_2017,visParam,"LZ_landcover_2017")

// download 
Export.image({
    
    image:LZ_landcover_2017,
    description: "LZ_landcover_2017",
    assetId: "LZ_landcover_2017",
    region:LZ,
    scale:10,
    maxPixels:1e13
});

You can go to “ Geoscience remote sensing data collection ”（www.gisrsdata.com） Get more data .

Reference resources

Gong P., et al., 2019. Stable classification with limited sample: transferring a 30-m resolution sample set collected in 2015 to mapping 10-m resolution global land cover in 2017,Science Bulletin

Li C, Gong P, Wang J, et al. The first all-season sample set for mapping global land cover with Landsat-8 data. Sci Bull 2017, 62, 508–515.

Finer Resolution Observation and Monitoring of Global Land Cover.http://data.ess.tsinghua.edu.cn/fromglc2017v1.html

PIE&GEE download ESA、Esri And FROM Three new 10 Meter land cover data code .https://zhuanlan.zhihu.com/p/424767735