SVM_2000.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "SVM_2000.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyO6atFQQmEzWNh6zs449Yo6",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/joostnjns/Google-Earth-Engine/blob/master/SVM_2000.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9hwQobNUzmjI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "var cost_svm = 70;\n",
        "\n",
        "// Query satellite image within bounding box\n",
        "var aoi = ee.Geometry.Rectangle([14.35, 26.95, 14.51, 27.12]);\n",
        "\n",
        "var collection00 = ee.ImageCollection('LANDSAT/LT05/C01/T1_SR').filterDate('2000-01-01', '2000-02-10').filterBounds(aoi).filterMetadata('IMAGE_QUALITY', 'equals', 9).select(['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7']);\n",
        "var image00 = collection00.sort('CLOUD_COVER').first();\n",
        "\n",
        "var collection05 = ee.ImageCollection('LANDSAT/LT05/C01/T1_SR').filterDate('2005-01-01', '2005-02-20').filterBounds(aoi).filterMetadata('IMAGE_QUALITY', 'equals', 9).select(['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7']);\n",
        "var image05 = collection05.sort('CLOUD_COVER').first();\n",
        "\n",
        "var collection10 = ee.ImageCollection('LANDSAT/LT05/C01/T1_SR').filterDate('2010-01-01', '2010-01-30').filterBounds(aoi).filterMetadata('IMAGE_QUALITY', 'equals', 9).select(['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7']);\n",
        "var image10 = collection10.sort('CLOUD_COVER').first();\n",
        "\n",
        "function calcMSAVI2(image){\n",
        "  var msavi2 = image.expression(\n",
        "  '((2 * NIR + 1 - sqrt(pow((2 * NIR + 1), 2) - 8 * (NIR - RED)) ) / 2)*1000',  //to be in same scale as spectral bands \n",
        "  {\n",
        "    'NIR': image.select('B4'), \n",
        "    'RED': image.select('B3')\n",
        "  }\n",
        "  );\n",
        "  image = image.addBands([msavi2]);\n",
        "  image = image.select(['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7','constant']).rename(['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7','msavi2']);\n",
        "  return image\n",
        "}\n",
        "\n",
        "var image00 = calcMSAVI2(image00);\n",
        "var image05 = calcMSAVI2(image05);\n",
        "var image10 = calcMSAVI2(image10);\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "// Define the visualization parameters.\n",
        "var vizParams1 = {\n",
        "  bands: ['B3','B2','B1'],\n",
        "  min: 0,\n",
        "  max: 5000,\n",
        "  gamma: [0.9,1.1,1]\n",
        "};\n",
        "\n",
        "// Visualize satellite image              \n",
        "Map.setCenter(14.42, 27.04, 13);\n",
        "\n",
        "Map.addLayer(image00.clip(aoi), vizParams1, 'true00');\n",
        "Map.addLayer(image05.clip(aoi), vizParams1, 'true05');\n",
        "//Map.addLayer(image10.clip(aoi), vizParams1, 'true10');\n",
        "\n",
        "\n",
        "//////////////////\n",
        "//   TRAINING   //\n",
        "//////////////////\n",
        "\n",
        "// Load training data and sample ground truth polygons\n",
        "var trainData = ee.FeatureCollection('users/jneujens/trainData2000');    // Load training data\n",
        "//Map.addLayer(trainData);\n",
        "\n",
        "var label = 'classname';     // Property containing land cover labels\n",
        "var groundTrue00 = image00.sampleRegions({     // Get training from polygons on image\n",
        "  collection: trainData,\n",
        "  properties: ['classvalue'],\n",
        "  scale: 30\n",
        "});\n",
        "\n",
        "var groundTrue05 = image05.sampleRegions({     // Get training from polygons on image\n",
        "  collection: trainData,\n",
        "  properties: ['classvalue'],\n",
        "  scale: 30\n",
        "});\n",
        "\n",
        "var groundTrue10 = image10.sampleRegions({     // Get training from polygons on image\n",
        "  collection: trainData,\n",
        "  properties: ['classvalue'],\n",
        "  scale: 30\n",
        "});\n",
        "\n",
        "// Split ground truth into training and test data\n",
        "var withRandom = groundTrue05.randomColumn('random');\n",
        "var split = 0.60;  // Roughly 70% training, 30% testing.\n",
        "var training = withRandom.filter(ee.Filter.lt('random', split));\n",
        "var testing = withRandom.filter(ee.Filter.gte('random', split));\n",
        "\n",
        "// Create an SVM classifier with custom parameters.\n",
        "var classifier00 = ee.Classifier.svm({\n",
        "  decisionProcedure: 'Voting',\n",
        "  svmType: 'C_SVC',\n",
        "  kernelType: 'LINEAR',\n",
        "  cost: 0.1,\n",
        "\n",
        "});\n",
        "\n",
        "var classifier05 = ee.Classifier.svm({\n",
        "  decisionProcedure: 'Voting',\n",
        "  svmType: 'C_SVC',\n",
        "  kernelType: 'LINEAR',\n",
        "  cost: 1500\n",
        "});\n",
        "\n",
        "var classifier10 = ee.Classifier.svm({\n",
        "  decisionProcedure: 'Voting',\n",
        "  svmType: 'C_SVC',\n",
        "  kernelType: 'LINEAR',\n",
        "  cost: 50\n",
        "});\n",
        "\n",
        "// Train the classifier.\n",
        "var trained00 = classifier00.train(groundTrue00, 'classvalue', ['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7', 'msavi2']);\n",
        "var trained05 = classifier05.train(training, 'classvalue', ['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7', 'msavi2']);\n",
        "//var trained10 = classifier10.train(groundTrue10, 'classvalue', ['B1', 'B2', 'B3', 'B4', 'B5', 'B6', 'B7', 'msavi2']);\n",
        "\n",
        "\n",
        "// Display results\n",
        "\n",
        "var sebha00 = image00.clip(aoi).classify(trained00);\n",
        "var sebha05 = image05.clip(aoi).classify(trained05);\n",
        "//var sebha10 = image10.clip(aoi).classify(trained10);\n",
        "\n",
        "var vizParams2 = {min: 0, max: 2, palette: ['grey', 'green','red']}\n",
        "\n",
        "// Display result\n",
        "\n",
        "Map.addLayer(sebha00, vizParams2, 'class00');\n",
        "Map.addLayer(sebha05, vizParams2, 'class05');\n",
        "//Map.addLayer(sebha10, vizParams2, 'class10');\n",
        "\n",
        "\n",
        "// // Export a cloud-optimized GeoTIFF.\n",
        "// Export.image.toDrive({\n",
        "//   image: sebha00,\n",
        "//   description: 'Result00',\n",
        "//   scale: 30,\n",
        "//   region: aoi,\n",
        "//   fileFormat: 'GeoTIFF',\n",
        "//   formatOptions: {\n",
        "//     cloudOptimized: true\n",
        "//   }\n",
        "// })\n",
        "\n",
        "// Export a cloud-optimized GeoTIFF.\n",
        "Export.image.toDrive({\n",
        "  image: sebha05,\n",
        "  description: 'Result05',\n",
        "  scale: 30,\n",
        "  region: aoi,\n",
        "  fileFormat: 'GeoTIFF',\n",
        "  formatOptions: {\n",
        "    cloudOptimized: true\n",
        "  }\n",
        "})\n",
        "\n",
        "// // Export a cloud-optimized GeoTIFF.\n",
        "// Export.image.toDrive({\n",
        "//   image: sebha10,\n",
        "//   description: 'Result10',\n",
        "//   scale: 30,\n",
        "//   region: aoi,\n",
        "//   fileFormat: 'GeoTIFF',\n",
        "//   formatOptions: {\n",
        "//     cloudOptimized: true\n",
        "//   }\n",
        "// })"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}