Video_Dehazing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chalo bhai chal gyi video \n"
     ]
    }
   ],
   "source": [
    "import cv2\n",
    "import math\n",
    "import numpy as np\n",
    "import sys\n",
    "\n",
    "def apply_mask(matrix, mask, fill_value):\n",
    "    #print(\"MATRIX=\", matrix)\n",
    "    #print(\"mask=\\n\" ,mask)\n",
    "    #print(\"fill value=\\n\", fill_value)\n",
    "                 \n",
    "    masked = np.ma.array(matrix, mask=mask, fill_value=fill_value)\n",
    "    #print('MASKED=',masked)\n",
    "    return masked.filled()\n",
    "\n",
    "def apply_threshold(matrix, low_value=255, high_value=255):\n",
    "    low_mask = matrix < low_value\n",
    "    #print(\"low mask=\",low_mask)\n",
    "    \n",
    "    matrix = apply_mask(matrix, low_mask, low_value)\n",
    "    #print('Low MASK->',low_mask,'\\nMatrix->',matrix)\n",
    "\n",
    "    high_mask = matrix > high_value\n",
    "    matrix = apply_mask(matrix, high_mask, high_value)\n",
    "\n",
    "    return matrix\n",
    "\n",
    "def simplest_cb(img, percent):\n",
    "    assert img.shape[2] == 3\n",
    "    assert percent > 0 and percent < 100\n",
    "    #print(\"shape of image = \", img.shape[2])\n",
    "\n",
    "    half_percent = percent / 200.0\n",
    "    #print('HALF PERCENT->',half_percent)\n",
    "\n",
    "    channels = cv2.split(img)\n",
    "    #print('Channels->\\n',channels)\n",
    "    #print('Shape->',channels[0].shape)\n",
    "    #print('Shape of channels->',len(channels[2]))\n",
    "\n",
    "    out_channels = []\n",
    "    for channel in channels:\n",
    "        assert len(channel.shape) == 2\n",
    "\n",
    "\t# find the low and high precentile values (based on the input percentile)\n",
    "        height, width = channel.shape\n",
    "        vec_size = width * height\n",
    "        flat = channel.reshape(vec_size)\n",
    "        #print('vec=',vec_size,'\\nFlat=',flat)\n",
    "        assert len(flat.shape) == 1\n",
    "\n",
    "        flat = np.sort(flat)\n",
    "\n",
    "        n_cols = flat.shape[0]\n",
    "        #print(\"Number of columns = \", n_cols)\n",
    "\n",
    "        low_val  = flat[math.floor(n_cols * half_percent)]\n",
    "        high_val = flat[math.ceil( n_cols * (1.0 - half_percent))]\n",
    "\n",
    "        #print(\"Lowval: \", low_val)\n",
    "        #print(\"Highval: \", high_val)\n",
    "        #print(flat[60])\n",
    "        #print(flat[11940])\n",
    "        \n",
    "\n",
    "        # saturate below the low percentile and above the high percentile\n",
    "        thresholded = apply_threshold(channel, low_val, high_val)\n",
    "        # scale the channel\n",
    "        normalized = cv2.normalize(thresholded, thresholded.copy(), 0, 255, cv2.NORM_MINMAX)\n",
    "        out_channels.append(normalized)\n",
    "\n",
    "    return cv2.merge(out_channels)\n",
    "\n",
    "if __name__ == '__main__':\n",
    "    #img = cv2.imread(sys.argv[1])\n",
    "    cap=cv2.VideoCapture('/media/dheeraj/9A26F0CB26F0AA01/WORK/github_repo/Dehazing/haze-videos/Whale.mov')\n",
    "    #img = cv2.imread('/home/dheeraj/Downloads/Whale.mov')\n",
    "    \n",
    "    fourcc = cv2.VideoWriter_fourcc(*'XVID')\n",
    "    frame_width=640\n",
    "    frame_height=480\n",
    "    save = cv2.VideoWriter('/media/dheeraj/9A26F0CB26F0AA01/WORK/github_repo/Dehazing/output.mov',fourcc, 20.0, (frame_width,frame_height))\n",
    "    \n",
    "    #out = cv2.VideoWriter('output.avi',fourcc, 20.0, (1920,1080))\n",
    "    while(cap.isOpened()):\n",
    "        ret, frame=cap.read()\n",
    "        if(ret== True):\n",
    "            \n",
    "            \n",
    "            frame=cv2.resize(frame, (frame_width, frame_height))\n",
    "            out = simplest_cb(frame, 1)\n",
    "            cv2.imshow(\"Before\",frame) \n",
    "            cv2.imshow(\"After\", out)\n",
    "            out = cv2.flip(out,0)\n",
    "            save.write(out)\n",
    "            key=cv2.waitKey(1)\n",
    "            if key==27:\n",
    "                break\n",
    "        else:\n",
    "            break\n",
    "\n",
    "    cap.release()\n",
    "    save.release()\n",
    "    cv2.destroyAllWindows()\n",
    "    \n",
    "    print(\"chalo bhai chal gyi video \")\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}