Create image chips that have consistent dimensions despite varying aspect ratios of input image region using Earth Engine and PIL

ee_image_chips.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/jdbcode/f3ebf33bfdf74a257b734a1ff61bc83b/ee_image_chips.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cwxd0NG5Nk_7"
      },
      "outputs": [],
      "source": [
        "import ee\n",
        "ee.Authenticate()\n",
        "ee.Initialize()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "FeatureCollection of cities"
      ],
      "metadata": {
        "id": "GOYgn3aSYS6E"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cities_fc = ee.FeatureCollection(\n",
        "        [ee.Feature(\n",
        "            ee.Geometry.Polygon(\n",
        "                [[[-86.30931396484377, 40.08385948719423],\n",
        "                  [-86.52080078125002, 39.87550140515563],\n",
        "                  [-86.53453369140627, 39.72145688768831],\n",
        "                  [-86.22966308593752, 39.44839825753037],\n",
        "                  [-85.81767578125002, 39.63901776816407],\n",
        "                  [-85.78197021484377, 39.88603977106767],\n",
        "                  [-85.89732666015627, 40.111172435524274]]]),\n",
        "            {\n",
        "              \"city_name\": \"city_0\"\n",
        "            }),\n",
        "        ee.Feature(\n",
        "            ee.Geometry.Polygon(\n",
        "                [[[-86.99870605468752, 40.51119785354961],\n",
        "                  [-87.00145263671877, 40.385790713738075],\n",
        "                  [-86.86961669921877, 40.29367679615214],\n",
        "                  [-86.66636962890627, 40.43598164994896]]]),\n",
        "            {\n",
        "              \"city_name\": \"city_1\"\n",
        "            }),\n",
        "        ee.Feature(\n",
        "            ee.Geometry.Polygon(\n",
        "                [[[-86.23515625000002, 40.52998873829066],\n",
        "                  [-86.19670410156252, 40.392066627828704],\n",
        "                  [-86.05113525390627, 40.35858832843792],\n",
        "                  [-86.03740234375002, 40.477778831614565],\n",
        "                  [-86.10881347656252, 40.55294822166985]]]),\n",
        "            {\n",
        "              \"city_name\": \"city_2\"\n",
        "            })]);"
      ],
      "metadata": {
        "id": "0EIiL4ipYB1v"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "An impervious surface image visualized as RGB"
      ],
      "metadata": {
        "id": "jBf6CD82YYtt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "img = ee.Image('USGS/NLCD_RELEASES/2019_REL/NLCD/2019').select('impervious')\n",
        "vis_img = img.visualize(min=0, max=100, palette=['white', 'yellow', 'brown', 'black'])"
      ],
      "metadata": {
        "id": "OK0lf8N4Ntok"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Get bounds and add buffer to each city"
      ],
      "metadata": {
        "id": "iXN41HiRYfuz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "city_names = cities_fc.aggregate_array('city_name').distinct()\n",
        "\n",
        "def city_box(city_name):\n",
        "    city_box = (cities_fc.filter(ee.Filter.eq('city_name', city_name))\n",
        "                  .geometry().bounds().buffer(1000, 1)) # 1 km buffer\n",
        "    return ee.Feature(city_box).set('city_name', city_name)\n",
        "\n",
        "city_boxes_fc = ee.FeatureCollection(city_names.map(city_box))"
      ],
      "metadata": {
        "id": "1Uvu1ufNYe1a"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Loop to download images"
      ],
      "metadata": {
        "id": "bal2V8KDaidh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import urllib.request\n",
        "from PIL import Image, ImageOps\n",
        "\n",
        "city_names_client = city_names.getInfo()\n",
        "\n",
        "def overlay_boundary(city_name):\n",
        "    city_features =  cities_fc.filter(ee.Filter.eq('city_name', city_name))\n",
        "    return vis_img.clipToCollection(city_features).unmask(0)\n",
        "\n",
        "def add_padding(img_file, target_size):\n",
        "    img = Image.open(img_file)\n",
        "    width, height = img.size\n",
        "\n",
        "    delta_w = target_size - width\n",
        "    delta_h = target_size - height\n",
        "\n",
        "    pad_left = delta_w // 2\n",
        "    pad_right = delta_w - pad_left\n",
        "    pad_top = delta_h // 2\n",
        "    pad_bottom = delta_h - pad_top\n",
        "\n",
        "    padded_img = ImageOps.expand(img, (pad_left, pad_top, pad_right, pad_bottom), fill=(0, 0, 0))\n",
        "\n",
        "    padded_img.save(f'{img_file[:-4]}_padded.png')\n",
        "\n",
        "for city_name in city_names_client:\n",
        "    print(f'Downloading: {city_name}')\n",
        "    image_to_download = overlay_boundary(city_name)\n",
        "    img_url = image_to_download.getThumbURL({\n",
        "        'dimensions': 1280,\n",
        "        'region': city_boxes_fc.filter(ee.Filter.eq('city_name', city_name)).geometry(),\n",
        "        'crs': 'EPSG:4326'\n",
        "        })\n",
        "\n",
        "    filename = f'{city_name}.png'\n",
        "    urllib.request.urlretrieve(img_url, filename)\n",
        "\n",
        "    add_padding(filename, 1280)"
      ],
      "metadata": {
        "id": "XBQ2JYGCamCV"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}