Resize-512x512.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Original PNG images are 1500x1500x3.  This will resize each image to 512x512x3 and put these files in\n",
    "a new directory."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "from PIL import Image\n",
    "import os \n",
    "from glob import glob\n",
    "import pathlib\n",
    "\n",
    "import numpy as np # linear algebra\n",
    "import matplotlib.pylab as plt\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Resize the 1500x1500 files to 512x512"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "INPUT_PATH_PNG = '/home/ubuntu/fastai/data/roads/mass_roads_png'\n",
    "OUTPUT_PATH_PNG_512 = '/home/ubuntu/fastai/data/roads/mass_roads_png_512'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check that the destination directories are in place."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mk_dirs(path, list_of_data_sets):\n",
    "    for p in list_of_data_sets:\n",
    "        sat = '{}/{}/sat'.format(path, p)\n",
    "        map = '{}/{}/map'.format(path, p)\n",
    "        pathlib.Path(sat).mkdir(parents=True, exist_ok=True)\n",
    "        pathlib.Path(map).mkdir(parents=True, exist_ok=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "mk_dirs(OUTPUT_PATH_PNG_512, ['train', 'valid', 'test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_ids(path, data_set):\n",
    "    ''' Look in the \"sat\" directory and get the filenames from here. The\n",
    "    same filenames appear in the \"map\" directory - no need to do it as well.'''\n",
    "    sat_dir = '{}/{}/sat/'.format(path, data_set)\n",
    "    file_dir = '{}*.png'.format(sat_dir)\n",
    "    files = glob(file_dir)\n",
    "    ids = [s[len(sat_dir):-4] for s in files]\n",
    "    return ids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['22978885_15', '25679230_15', '22678870_15', '17128885_15', '23428960_15']"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_ids(INPUT_PATH_PNG, 'train')[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['24479170_15', '17878735_15', '18178780_15', '10978735_15', '22978990_15']"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_ids(INPUT_PATH_PNG, 'valid')[:5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Resize the TRAINing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def resize_img(fn, data_set, subdir):\n",
    "    input_file = '{}/{}/{}/{}.png'.format(INPUT_PATH_PNG, data_set, subdir, fn)\n",
    "    output_file = '{}/{}/{}/{}.png'.format(OUTPUT_PATH_PNG_512, data_set, subdir, fn)\n",
    "    return(input_file, output_file)\n",
    "\n",
    "def resize_img_set(data_set, fn, size):\n",
    "    for subdir in ['map', 'sat']:\n",
    "        input_file, output_file = resize_img(fn, data_set, subdir)\n",
    "        Image.open(input_file).resize((size,size)).save(output_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_ids = get_ids(INPUT_PATH_PNG, 'train')\n",
    "for fn in train_ids:\n",
    "    resize_img_set('train', fn, 512)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Resize the VALIDation data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "valid_ids = get_ids(INPUT_PATH_PNG, 'valid')\n",
    "for fn in valid_ids:\n",
    "    resize_img_set('valid', fn, 512)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Resize TEST data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_ids = get_ids(INPUT_PATH_PNG, 'test')\n",
    "for fn in test_ids:\n",
    "    resize_img_set('test', fn, 512)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Display a single image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_filenames(path, data_set):\n",
    "    ''' Look in the \"sat\" directory and get the filenames from here. The\n",
    "    same filenames appear in the \"map\" directory - no need to do it as well.'''\n",
    "    sat_dir = '{}/{}/sat/'.format(path, data_set)\n",
    "    sat_file_dir = '{}*.png'.format(sat_dir)\n",
    "    sat_files = glob(sat_file_dir)\n",
    "    map_dir = '{}/{}/map/'.format(path, data_set)\n",
    "    map_file_dir = '{}*.png'.format(map_dir)\n",
    "    map_files = glob(map_file_dir)\n",
    "    return (sat_files, map_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['/home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat/22978885_15.png',\n",
       " '/home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat/25679230_15.png',\n",
       " '/home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat/22678870_15.png',\n",
       " '/home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat/17128885_15.png',\n",
       " '/home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat/23428960_15.png']"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_sat_fns, train_map_fns = get_filenames(OUTPUT_PATH_PNG_512, 'train')\n",
    "train_sat_fns[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1108\r\n"
     ]
    }
   ],
   "source": [
    "!ls /home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat | wc -l\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_files = !ls /home/ubuntu/fastai/data/roads/mass_roads_png_512/train/sat | wc -l\n",
    "nf = int(num_files[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Image shape: (512, 512, 3) | image type: uint8 | mask shape: (512, 512) | mask type: uint8\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f1c773cd240>"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fm = random.choice(train_map_fns)\n",
    "fs = random.choice(train_sat_fns)\n",
    "\n",
    "mask = cv2.imread(fm,0)\n",
    "img = cv2.imread(fs)\n",
    "\n",
    "img_masked = cv2.bitwise_and(img, img, mask=mask)\n",
    "\n",
    "print(\"Image shape: {} | image type: {} | mask shape: {} | mask type: {}\".format(img.shape, img.dtype, mask.shape, mask.dtype) )\n",
    "\n",
    "plt.subplot(131)\n",
    "plt.imshow(img)\n",
    "plt.subplot(132)\n",
    "plt.imshow(mask)\n",
    "plt.subplot(133)\n",
    "plt.imshow(img_masked)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}