01_manipulate_JSON_label_files.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "01_manipulate JSON label files.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "mount_file_id": "1jV4cZowukS3EpWdXuwpMtsIArV1nLGXB",
      "authorship_tag": "ABX9TyMJzlq53AYrzQnZtvFs0RjP",
      "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/Seyed-Ali-Ahmadi/Building-Damage-Assessment/blob/master/01_manipulate_JSON_label_files.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sMYCp5dUKYmH",
        "colab_type": "text"
      },
      "source": [
        "# Section 1: Manipulation\n",
        "\n",
        "In this section I will try to create building masks for a single image."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rQzHhJEp7DNo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import json\n",
        "import re\n",
        "from skimage.draw import polygon\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JmGD1Dnq800R",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "path = '/content/drive/My Drive/subset_1/guatemala-volcano_00000000_post_disaster.json'\n",
        "# METHOD 1\n",
        "with open(path) as f:\n",
        "  data = json.load(f)\n",
        "\n",
        "# METHOD 2\n",
        "# Alternative solution to read the JSON file:\n",
        "# Opening JSON file \n",
        "f = open(path,) \n",
        "# returns JSON object as  \n",
        "# a dictionary \n",
        "data = json.load(f) "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BjMOWaGn-jxj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 207
        },
        "outputId": "09a417e4-0ec6-44a1-9d40-df57d7d33d30"
      },
      "source": [
        "print(type(data))\n",
        "print('---------------------')\n",
        "print(data.keys())\n",
        "print('---------------------')\n",
        "for key in data.keys():\n",
        "  print('The key whose keys are being displayed is \"' + key + '\"')\n",
        "  print(data[key].keys())\n",
        "  print('............')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'dict'>\n",
            "---------------------\n",
            "dict_keys(['features', 'metadata'])\n",
            "---------------------\n",
            "The key whose keys are being displayed is \"features\"\n",
            "dict_keys(['lng_lat', 'xy'])\n",
            "............\n",
            "The key whose keys are being displayed is \"metadata\"\n",
            "dict_keys(['sensor', 'provider_asset_type', 'gsd', 'capture_date', 'off_nadir_angle', 'pan_resolution', 'sun_azimuth', 'sun_elevation', 'target_azimuth', 'disaster', 'disaster_type', 'catalog_id', 'original_width', 'original_height', 'width', 'height', 'id', 'img_name'])\n",
            "............\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LJ8CMyzg9-5W",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8e06f9af-37aa-4a3c-c3e5-91c953c296e7"
      },
      "source": [
        "print('There are << ' + str(len(data['features']['xy'])) + ' >> buildings in this image.')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "There are << 10 >> buildings in this image.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FG4s5CT6BYCO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        },
        "outputId": "471fea87-1f3d-45b2-fdbd-7a160a42d5d5"
      },
      "source": [
        "building_mask = np.zeros((1024, 1024))\n",
        "\n",
        "for building in data['features']['xy']:\n",
        "  \n",
        "  vertices = building['wkt']\n",
        "  vertices = vertices.partition('POLYGON ((')[2].partition('))')[0].split(', ')\n",
        "  n_vertices = len(vertices)\n",
        "\n",
        "  print('This is a < ' + \n",
        "        building['properties']['feature_type'] + \n",
        "        ' > of type << ' + \n",
        "        building['properties']['subtype'] + ' >> with ' + \n",
        "        str(n_vertices) + ' vertices. ')\n",
        "\n",
        "  rows = []\n",
        "  cols = []\n",
        "  for vertex in vertices:\n",
        "    cols.append(float(vertex.split(' ')[0]))\n",
        "    rows.append(float(vertex.split(' ')[1]))\n",
        "\n",
        "  rr, cc = polygon(rows, cols, (1024, 1024))\n",
        "  \n",
        "  building_mask[rr, cc] = 1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "This is a < building > of type << no-damage >> with 9 vertices. \n",
            "This is a < building > of type << no-damage >> with 15 vertices. \n",
            "This is a < building > of type << no-damage >> with 9 vertices. \n",
            "This is a < building > of type << no-damage >> with 6 vertices. \n",
            "This is a < building > of type << no-damage >> with 10 vertices. \n",
            "This is a < building > of type << no-damage >> with 7 vertices. \n",
            "This is a < building > of type << no-damage >> with 6 vertices. \n",
            "This is a < building > of type << no-damage >> with 5 vertices. \n",
            "This is a < building > of type << no-damage >> with 4 vertices. \n",
            "This is a < building > of type << no-damage >> with 5 vertices. \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mhFUC2QTJywn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 487
        },
        "outputId": "b7d5b848-d060-46a4-c4bb-b41e0141506f"
      },
      "source": [
        "plt.figure(figsize=(8, 8))\n",
        "plt.imshow(building_mask, cmap='gray')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DUoJT6gHKopd",
        "colab_type": "text"
      },
      "source": [
        "# Section 2: Automation\n",
        "\n",
        "In this section I will try to make the above code more efficient such that it gets a directory as input and creates the mask images as outputs based on their GeoJSON label files.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "osiknKCdK56G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import re\n",
        "import json\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from skimage.draw import polygon\n",
        "from skimage.io import imread\n",
        "\n",
        "\n",
        "rootDir = '/content/drive/My Drive/subset_labels/'\n",
        "jsonFiles = os.listdir(rootDir)\n",
        "\n",
        "imageDir = '/content/drive/My Drive/subset_images/'\n",
        "\n",
        "for jsonFile in jsonFiles:\n",
        "  f = open(rootDir + jsonFile,)\n",
        "  data = json.load(f)\n",
        "\n",
        "  # Create a new building mask.\n",
        "  building_mask = np.zeros((1024, 1024))\n",
        "  bounding_box = np.zeros((1024, 1024))\n",
        "\n",
        "  # Read the corresponding image of the same JSON file.\n",
        "  image = imread(imageDir + jsonFile[:-4] + 'png')\n",
        "\n",
        "  # Loop through buildings in each image.\n",
        "  for building in data['features']['xy']:\n",
        "    \n",
        "    # Extract x-y coordinate of each vertex from decoding the json pattern.\n",
        "    vertices = building['wkt'].partition('POLYGON ((')[2].partition('))')[0].split(', ')\n",
        "    n_vertices = len(vertices)\n",
        "\n",
        "    rows = []\n",
        "    cols = []\n",
        "    for vertex in vertices:\n",
        "      cols.append(float(vertex.split(' ')[0]))\n",
        "      rows.append(float(vertex.split(' ')[1]))\n",
        "\n",
        "    # Fill the location of each building.\n",
        "    rr, cc = polygon(rows, cols, (1024, 1024))\n",
        "    building_mask[rr, cc] = 1\n",
        "    # Fill the location of its bounding box for further use.\n",
        "    br, bc = polygon([min(rows), min(rows), max(rows), max(rows), min(rows)],\n",
        "                     [min(cols), max(cols), max(cols), min(cols), min(cols)],\n",
        "                     (1024, 1024))    \n",
        "    bounding_box[br, bc] = 1\n",
        "\n",
        "    mask_box = building_mask + bounding_box\n",
        "    building_crop = image[int(np.floor(min(rows))) - 10:\n",
        "                          int(np.ceil(max(rows))) + 10,\n",
        "                          int(np.floor(min(cols))) - 10:\n",
        "                          int(np.ceil(max(cols))) + 10, :]\n",
        "    mask_crop = mask_box[int(np.floor(min(rows))) - 10:\n",
        "                         int(np.ceil(max(rows))) + 10,\n",
        "                         int(np.floor(min(cols))) - 10:\n",
        "                         int(np.ceil(max(cols))) + 10]\n",
        "\n",
        "    plt.figure(figsize=(15, 12))\n",
        "    plt.subplot(121), plt.imshow(mask_crop)\n",
        "    plt.subplot(122), plt.imshow(building_crop)\n",
        "    plt.show()\n",
        "\n",
        "  \n",
        "  # plt.figure(figsize=(15, 12))\n",
        "  # plt.subplot(131), plt.imshow(mask_box, cmap='jet')\n",
        "  # plt.subplot(132), plt.imshow(mask_crop)\n",
        "  # plt.subplot(133), plt.imshow(building_crop)\n",
        "  # plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}