Module 4 exercise

[edythneska]

Hello,
Please I have been able to upload my zipfile as an asset into GEE and imported but no scripts appear for me to be able to run it. Attached is the screenshot.

Thank you

[JuliePeeling]

Hi Edith,

There is a link to the script, '03_Landcover_Classification' in step 9 of the the exercise page that directs you to the script. Also, you can click the link here, and it should also direct you to the script that you can copy and paste into the Google Earth Engine code editor. Please let us know of any further issues.

Best,
Julie

[edythneska]

Hello Julie,

Thanks for sharing the script with me. But I noticed the link to the script wasn't active.

Thanks once again

[JuliePeeling]

I am unsure why the link is not active for you right now, but I will paste the script below while we work on troubleshooting this. Thank you for letting us know.

Best,
Julie

[JuliePeeling]

// Note: table comes from the shapefile given to you from the UF team
// ... geometry is drawn by you.
// Note: Make sure the geometry covers the points in the classification table

/** ############################### DEFAULTS ############################### **/

// REGION OF INTEREST
var REG_GH = geometry;

// var label = 'LCODE'; // This is the field containing the land cover class coded as a number.

// SET IMAGE SOURCES
var IMG_L5 = 'LANDSAT/LT05/C01/T1_SR'; // FOR 2000 - 2012
var IMG_L7 = 'LANDSAT/LE07/C01/T1_SR'; // FOR 2000 - 2012
var IMG_L8 = 'LANDSAT/LC08/C01/T1_SR'; // FOR 2017 - Present

// SET IMAGE ATTRIBUTES
var YEAR1 = 2019; // Note: we keep this the same as we have data from just one year.
var YEAR2 = 2019;

// SET DATES OF INTEREST
var ST_DATE = YEAR1+'-01-01'; // Note, the variable YEAR is defined above
var EN_DATE = YEAR2+'-12-31'; // Note, the variable YEAR is defined above

/** ############################### DEFINE FUNCTIONS ############################### **/

// RENAMING BANDS [This is a 'helper function' that will allow us to work across all Landsat functions, see in-class discussion]
// This function gets us images for out start and end dates, clips them to the region, applies cloud masks, and renames bands.
// ...Notice that this is exactly the same as for exercise 1, just a few more lines to help regularize band names.

var getLSAT = function(imgColl,REG_GH,stDate,enDate) {
var IMG_SR=ee.ImageCollection(imgColl)
.filterDate(stDate, enDate) // Filter images to dates we require
.filterBounds(REG_GH) // Filter image footprints by region
.map(function(image){return image.clip(REG_GH)}) // Clip images to region
.map(cloudMask) // Mask clouds using the 'cloudMask' function
// NEW: Rename bands to harmonize across sensors
.select(['B1','B2','B3','B4','B5','B6','B7','pixel_qa']); // Note: These are standard band names in all Landsat products
var bNames = ['blu','grn','red','nir','sw1','t1','sw2','pixel_qa']; // ... we create 'new' band names for Landsat 5 and 7
if (imgColl == 'LANDSAT/LC08/C01/T1_SR') // ... if we're looking at Landsat 8 though
{
bNames = ['ubl','blu','grn','red','nir','sw1','sw2','pixel_qa']; // ... we add in the corresponding new names (note ultra-blue, and changes in the thermal and swir2 bands)
}
var IMG_RN=IMG_SR.select(['B1','B2','B3','B4','B5','B6','B7','pixel_qa'],bNames); // ... we finally extract the only ones we want and rename them to the new names.
return IMG_RN.select(['blu','grn','red','nir','sw1','sw2','pixel_qa']); // Return the image collection with the renamed bands
};

// MASKING CLOUDS
var cloudMask = function(image) {
var qa = image.select('pixel_qa');
var cloud = qa.bitwiseAnd(1 << 5)
.and(qa.bitwiseAnd(1 << 7))
.or(qa.bitwiseAnd(1 << 3));
var mask2 = image.mask().reduce(ee.Reducer.min());
return image.updateMask(cloud.not()).updateMask(mask2);
};

// GET INDICES (see other indices at the end of the script)
var getNDVI = function(image){
var ind = image.expression('(nir-red)/(nir+red)',{'nir':image.select('nir'),'red':image.select('red')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getEVI = function(image){
var ind = image.expression('2.5 * ((NIR - Red) / (NIR + 6.0 * Red - 7.5 * Blue + 1))', {'NIR': image.select('nir'),'Red': image.select('red'),'Blue': image.select('blu')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getMSAVI = function(image){
var ind = image.expression('MSAVI = (1/2)(2(NIR+1)-sqrt((2*NIR+1)*2-8(NIR-Red)))',{'NIR': image.select('nir'),'Red': image.select('red')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getSAVI = function(image){
var ind = image.expression('SAVI = ((NIR - Red) / (NIR + Red + 0.5)) * (1 + 0.5)',{'NIR': image.select('nir'),'Red': image.select('red')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getVARI = function(image){
var ind = image.expression('VARI = (Green - Red)/ (Green + Red - Blue)',{'Green': image.select('grn'),'Red': image.select('red'),'Blue': image.select('blu') });
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getNDSI = function(image){
var ind = image.expression('NDSI = (Green - SWIR) / (Green + SWIR)',{'Green': image.select('grn'),'SWIR': image.select('sw1')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

var getMSI = function(image){
var ind = image.expression('MSI = (NIR - SWIR) / (NIR+ SWIR)',{'NIR': image.select('nir'),'SWIR': image.select('sw1')});
return ind.copyProperties({source: image,properties: ['system:time_start']});
};

/** ############################### PROCESS ############################### **/

// LET US GET ALL SATELLITE DATA THAT IS AVAILABLE

// var IMG_L5 = 'LANDSAT/LT05/C01/T1_SR'; // FOR 2000 - 2012 // For reference from above
// var IMG_L7 = 'LANDSAT/LE07/C01/T1_SR'; // FOR 2000 - 2012
// var IMG_L8 = 'LANDSAT/LC08/C01/T1_SR'; // FOR 2017 - Present

var IMG_LT5 = getLSAT(IMG_L5, REG_GH, ST_DATE, EN_DATE); // Landsat 5 (See details in last exercise)
var IMG_LT7 = getLSAT(IMG_L7, REG_GH, ST_DATE, EN_DATE); // Landsat 7 (See details in last exercise)
var IMG_LT8 = getLSAT(IMG_L8, REG_GH, ST_DATE, EN_DATE); // Landsat 8 (See details in last exercise)

// COMBINE ALL IMAGES INTO ONE STACK

var IMG_LTX = IMG_LT5.merge(IMG_LT7); // First, merge L5 and L7 into one...
var IMG_LTX = IMG_LTX.merge(IMG_LT8); // Then, merge L8 into this one
var IMG_LTX = IMG_LTX.sort('system:time_start');
print('Landsat collection for this model'); // Check total number of images available
print(IMG_LTX);
var LTX_MED = IMG_LTX.median().select(['blu','grn','red','nir','sw1','sw2']);// Get median of all bands

// GENERATE INDEX STACKS
var IMG_IND = IMG_LTX.map(getNDVI).median(); // Generate stack of NDVIs, take median
var IMG_IND = IMG_IND.addBands(IMG_LTX.map(getEVI).median()); // Generate stack of EVIs, take median, add to median NDVI as band
var IMG_IND = IMG_IND.addBands(IMG_LTX.map(getMSI).median()); // ...so on and so forth,
var IMG_IND = IMG_IND.addBands(IMG_LTX.map(getSAVI).median());
var IMG_IND = IMG_IND.addBands(IMG_LTX.map(getMSAVI).median());
var IMG_IND = IMG_IND.addBands(IMG_LTX.map(getVARI).median());

/** ############################### CLASSIFY ############################### **/
var table = table2;
var label='LCODE2';

// add a random column (by default named 'random'), will contain random numbers between 0-1
table = table.randomColumn();
// split in a training (80%) and validation (20%)
var train_samp = table.filter(ee.Filter.gt('random',0.2)); // 80% training
var test_samp = table.filter(ee.Filter.lte('random',0.2)); // 20% testing

// Select the image you want to classify with
//var IMG_CLS = IMG_IND;
var IMG_CLS = LTX_MED;

// SAMPLE FROM GROUND-TRUTH LOCATIONS
var training = IMG_CLS.sampleRegions({
collection: train_samp,
properties: [label],
scale: 30
});

print('Training dataset');
print(training);

var testing = IMG_CLS.sampleRegions({
collection: test_samp,
properties: [label],
scale: 30
});
print('Testing dataset');
print(testing);

// BUILD CLASSIFIER (see Classifier in the Docs pane on the left)
//var classifier = ee.Classifier.minimumDistance().train(training,label);
//var classifier = ee.Classifier.libsvm().train(training,label);
//var classifier = ee.Classifier.smileCart().train(training,label);
var classifier = ee.Classifier.smileRandomForest(10).train(training,label);

// Get a confusion matrix representing resubstitution accuracy.
var trainAccuracy = classifier.confusionMatrix();
print('Resubstitution error matrix: ', trainAccuracy); // Training accuracy matrix
print('Training overall accuracy: ', trainAccuracy.accuracy()); // Training accuracy

// Overlay the points on the imagery to get training.
var classified = IMG_CLS.classify(classifier,'Classified'); // Classify the input imagery.

// VALIDATE
var testclass = testing.classify(classifier,'Classified'); // Apply the classifier on the testing data.
print(testclass);
var testAccuracy = testclass.errorMatrix(label, 'Classified'); // Compute confusion matrix
print('Validation error matrix: ', testAccuracy); // Testing confusion matrix
print('Validation overall accuracy: ', testAccuracy.accuracy()); // Validation accuracy

/** ############################### VISUALIZE ############################### **/

// Visualization parameters are now common for Landsat 5, 7, and 8
var visParams = {
bands: ['nir', 'red', 'grn'],
min: 0, //Value to map to 0
max: 3000, //Value to map to 255
gamma: 1.4, //gamma correction factor controls the brightness
};

var visNDVI = {
min: -0.2, //Value to map to 0
max: 1.0, //Value to map to 255
palette: ['blue', 'white', 'green'],
};

/**
var palette = [
'FFFFFF', //0 Nodata?
'FFFF33', //1 Agriculture
'CCCCCC', //2 Barren
'006633', //3 Forest
'CC0000', //4 Urban
'0000CC', //5 Water
];
**/

var palette1 = [
'FFFFFF', // 0 Nodata?
'9aa4a6', // 1 Bare
'fbe839', // 2 Cropland
'e5071d', // 3 Built up
'0c8000', // 4 Forest
'a3d06c', // 5 Herbaceous
'b6ab08', // 6 Shrub
'030edb', // 7 Water
];

var palette2 = [
'FFFFFF', // 0 Nodata?
'BEC5C6',//Bare_Bare soil
'9AA4A6',//Bare_Rocky land
'C5C4B3',//Bare_Sandy area
'FBE839',//Cropland_Agriculture
'E5D005',//Cropland_Agriculture in shallows and recession
'C4B204',//Cropland_Cropland and fallow with oil palm
'C7A309',//Cropland_Irrigated agriculture
'AB0515',//Developed_Open mine
'E5071D',//Developed_Settlements
'546113',//Forest_Degraded forest
'0C8000',//Forest_Forest
'1E625C',//Forest_Gallery and riparian forest
'2B8D84',//Forest_Mangrove
'10AC00',//Forest_Plantation
'18FE00',//Forest_Swamp forest
'CAE4AA',//Herbaceous_Bowe
'A3D06C',//Herbaceous_Herbaceous savanna
'7DB43A',//Herbaceous_Steppe
'FCF7B2',//Shrub_Sahelian shortgrass savana
'F6D238',//Shrub_Savanna
'EFF509',//Shrub_Shrubland
'CDD208',//Shrub_Thicket
'A19607',//Shrub_Woodland
'030EDB',//Water bodies_Water bodies
'0199DD',//Water bodies_Wetland floodplain
];

Map.centerObject(REG_GH, 9); //Center the map at the region of interest
Map.addLayer(table,{},'Sampling ponits'); // Add sampling shapefile to the map
Map.addLayer(IMG_LTX.median(),visParams,'Landsat'); // Add image to the map
//Map.addLayer(classified,{min: 0, max: 7, palette: palette1},'Land cover 1');
Map.addLayer(classified,{min: 0, max: 25, palette: palette2},'Land cover 2');

[edythneska]

Copied. Thank you