Detector de Automóviles - (HOG, HOGP + SVM) - CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import cv2\n",
    "from skimage.feature import hog\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "import glob\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "vehiculos_n_archivos = glob.glob('dataset/vehicles/**/*.png')\n",
    "novehiculos_n_archivos = glob.glob('dataset/non-vehicles/**/*.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de imágenes de automóviles:  8792\n",
      "Cantidad de imágenes de no automóviles:  8968\n"
     ]
    }
   ],
   "source": [
    "print \"Cantidad de imágenes de automóviles: \", len(vehiculos_n_archivos)\n",
    "print \"Cantidad de imágenes de no automóviles: \", len(novehiculos_n_archivos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'No vehiculo')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img_vehiculo = cv2.imread(vehiculos_n_archivos[3])\n",
    "img_novehiculo = cv2.imread(novehiculos_n_archivos[6])\n",
    "\n",
    "fig, (vehiculo_plt, novehiculo_plt) = plt.subplots(1, 2)\n",
    "\n",
    "vehiculo_plt.imshow(cv2.cvtColor(img_vehiculo, cv2.COLOR_BGR2RGB))\n",
    "vehiculo_plt.set_title('Vehiculo')\n",
    "\n",
    "novehiculo_plt.imshow(cv2.cvtColor(img_novehiculo, cv2.COLOR_BGR2RGB))\n",
    "novehiculo_plt.set_title('No vehiculo')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Extractor de Características: HOG (Histogram Oriented Gradient)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "carac_hog_vehiculo, img_vehiculo_hog = hog(img_vehiculo[:,:,0], orientations = 9, pixels_per_cell=(8, 8),\n",
    "                    cells_per_block=(2, 2), transform_sqrt=True, visualise=True, feature_vector=True)\n",
    "\n",
    "carac_hog_novehiculo, img_novehiculo_hog = hog(img_novehiculo[:,:,0], orientations = 9, pixels_per_cell=(8, 8),\n",
    "                    cells_per_block=(2, 2), transform_sqrt=True, visualise=True, feature_vector=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'No Vehiculo: HOG')"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig2, (hog_vehiculo_plt, hog_novehiculo_plt) = plt.subplots(1, 2)\n",
    "\n",
    "hog_vehiculo_plt.imshow(img_vehiculo_hog, cmap='gray')\n",
    "hog_vehiculo_plt.set_title('Vehiculo: HOG')\n",
    "\n",
    "hog_novehiculo_plt.imshow(img_novehiculo_hog, cmap='gray')\n",
    "hog_novehiculo_plt.set_title('No Vehiculo: HOG')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dimensiones de la imagen:  (64, 64, 3)\n"
     ]
    }
   ],
   "source": [
    "print \"Dimensiones de la imagen: \", img_vehiculo.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def obtener_vector_caracteristicas_all(imgs, orient, pix_per_cell, cell_per_block, hog_channel, color_space=\"RGB\"):\n",
    "    '''\n",
    "    Funcion que retorna el vector de caracteristicas de HOG de una lista de imagenes\n",
    "    '''\n",
    "    caracteristicas = []\n",
    "    for file_img in imgs:\n",
    "        img = mpimg.imread(file_img)\n",
    "        \n",
    "        # Para cambiar el color_space de la imagen (si asi se lo desea) --> Agregar mas para realizar el testeo\n",
    "        if color_space != \"RGB\":\n",
    "            if color_space == 'HLS':\n",
    "                carac_img = cv2.cvtColor(img, cv2.COLOR_RGB2HLS)\n",
    "            elif color_space == 'HSV':\n",
    "                carac_img = cv2.cvtColor(img, cv2.COLOR_RGB2HSV)\n",
    "        else:\n",
    "            carac_img = np.copy(img)\n",
    "        \n",
    "        # si hog_chanell es 3 (quiere decir todos los canales)\n",
    "        if hog_channel == 3:\n",
    "            hog_caracteristicas =[]\n",
    "            for c in range(carac_img.shape[2]):\n",
    "                hog_carac_img = hog(carac_img[:,:,c], orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),\n",
    "                                  cells_per_block=(cell_per_block, cell_per_block), transform_sqrt=True, visualise=False, feature_vector=True)\n",
    "                hog_caracteristicas.append(hog_carac_img)\n",
    "            hog_caracteristicas = np.ravel(hog_caracteristicas)\n",
    "        else:\n",
    "            hog_caracteristicas = hog(carac_img[:,:,hog_channel], orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),\n",
    "                                  cells_per_block=(cell_per_block, cell_per_block), transform_sqrt=True, visualise=False, feature_vector=True)\n",
    "        \n",
    "        caracteristicas.append(hog_caracteristicas)\n",
    "    \n",
    "    return caracteristicas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def obtener_vector_caracteristicas_img(imagen, orient, pix_per_cell, cell_per_block, hog_channel, color_space=\"RGB\"):\n",
    "    '''\n",
    "    Funcion que retorna el vector de caracteristicas de HOG de una lista de imagenes\n",
    "    '''\n",
    "    caracteristicas = []\n",
    "\n",
    "    img = np.copy(imagen)\n",
    "\n",
    "    # Para cambiar el color_space de la imagen (si asi se lo desea) --> Agregar mas para realizar el testeo\n",
    "    if color_space != \"RGB\":\n",
    "        if color_space == 'HLS':\n",
    "            carac_img = cv2.cvtColor(img, cv2.COLOR_RGB2HLS)\n",
    "        elif color_space == 'HSV':\n",
    "            carac_img = cv2.cvtColor(img, cv2.COLOR_RGB2HSV)\n",
    "    else:\n",
    "        carac_img = np.copy(img)\n",
    "\n",
    "    # si hog_chanell es 3 (quiere decir todos los canales)\n",
    "    if hog_channel == 3:\n",
    "        hog_caracteristicas =[]\n",
    "        for c in range(carac_img.shape[2]):\n",
    "            hog_carac_img = hog(carac_img[:,:,c], orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),\n",
    "                              cells_per_block=(cell_per_block, cell_per_block), transform_sqrt=True, visualise=False, feature_vector=True)\n",
    "            hog_caracteristicas.append(hog_carac_img)\n",
    "        hog_caracteristicas = np.ravel(hog_caracteristicas)\n",
    "    else:\n",
    "        hog_caracteristicas = hog(carac_img[:,:,hog_channel], orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell),\n",
    "                              cells_per_block=(cell_per_block, cell_per_block), transform_sqrt=True, visualise=False, feature_vector=True)\n",
    "\n",
    "    caracteristicas.append(hog_caracteristicas)\n",
    "    \n",
    "    return caracteristicas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Entrenamiento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "hog_caracteristicas_vehiculos = obtener_vector_caracteristicas_all(vehiculos_n_archivos, \n",
    "                                                                orient=9,\n",
    "                                                                pix_per_cell=8,\n",
    "                                                                cell_per_block=2,\n",
    "                                                                hog_channel=3,\n",
    "                                                                color_space=\"RGB\")\n",
    "\n",
    "hog_caracteristicas_no_vehiculos = obtener_vector_caracteristicas_all(novehiculos_n_archivos, \n",
    "                                                                orient=9,\n",
    "                                                                pix_per_cell=8,\n",
    "                                                                cell_per_block=2,\n",
    "                                                                hog_channel=3,\n",
    "                                                                color_space=\"RGB\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Estandarizacion de Caracteristicas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = np.vstack((hog_caracteristicas_vehiculos, hog_caracteristicas_no_vehiculos)).astype(np.float64)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler_std = StandardScaler()\n",
    "\n",
    "# Calcula la media y el desvio estándar que se utilizaran para escalar posteriormente.\n",
    "X_scaler = scaler_std.fit(X)\n",
    "# Realiza la estandarizacion centrando y escalando\n",
    "X_scaled = X_scaler.transform(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = np.hstack((np.ones(len(hog_caracteristicas_vehiculos)), np.zeros(len(hog_caracteristicas_no_vehiculos))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_svc = svm.LinearSVC()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "     intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "     multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
       "     verbose=0)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_svc.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9664977477477478"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_svc.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicción:  [0. 1. 1. 0. 1. 0. 1. 1. 0. 0. 0. 1. 0. 0. 0. 0. 1. 1. 0. 0.]\n",
      "Clase:       [0. 1. 1. 0. 1. 0. 1. 1. 0. 0. 0. 1. 0. 0. 0. 1. 1. 1. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "print 'Predicción: ', model_svc.predict(X_test[20:40])\n",
    "print 'Clase:      ', y_test[20:40]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_test[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Detección de vehiculos en una imagen"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ventana_deslizante(img, \n",
    "                 x_ini_fin=[None, None], \n",
    "                 y_ini_fin=[None, None], \n",
    "                 xy_ventana=(64, 64), \n",
    "                 xy_superposicion=(0.5, 0.5)):\n",
    "       \n",
    "    x_ancho = x_ini_fin[1] - x_ini_fin[0]\n",
    "    y_ancho = y_ini_fin[1] - y_ini_fin[0]\n",
    "    \n",
    "    # Cantidad de pixeles por paso\n",
    "    nx_pix_paso = np.int(xy_ventana[0] * (1 - xy_superposicion[0]))\n",
    "    ny_pix_paso = np.int(xy_ventana[1] * (1 - xy_superposicion[1]))\n",
    "    \n",
    "    # Calcula la cantidad de ventanas en x/y\n",
    "    nx_buffer = np.int(xy_ventana[0] * (xy_superposicion[0]))\n",
    "    ny_buffer = np.int(xy_ventana[1] * (xy_superposicion[1]))\n",
    "    nx_ventanas = np.int((x_ancho - nx_buffer) / nx_pix_paso) \n",
    "    ny_ventanas = np.int((y_ancho - ny_buffer) / ny_pix_paso) \n",
    "    \n",
    "    # Lista de posiciones de ventanas\n",
    "    lista_ventanas = []\n",
    "    \n",
    "    # Calcular las posiciones de la ventanas\n",
    "    for ys in range(ny_ventanas):\n",
    "        for xs in range(nx_ventanas):\n",
    "            ini_x = xs*nx_pix_paso + x_ini_fin[0]\n",
    "            fin_x = ini_x + xy_ventana[0]\n",
    "            ini_y = ys*ny_pix_paso + y_ini_fin[0]\n",
    "            fin_y = ini_y + xy_ventana[1]\n",
    "            \n",
    "            lista_ventanas.append(((ini_x, ini_y), (fin_x, fin_y)))\n",
    "\n",
    "    return lista_ventanas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dibujar_rectangulo(img, rectangulos):\n",
    "    img_rec = np.copy(img)\n",
    "    color = (255,255,0)\n",
    "    ancho = 4\n",
    "    \n",
    "    for rec in rectangulos:\n",
    "        cv2.rectangle(img_rec, rec[0], rec[1], color, ancho)\n",
    "        \n",
    "    return img_rec\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def buscar_vehiculos(img, ventanas):\n",
    "    \n",
    "    for ventana in ventanas:\n",
    "        \n",
    "        ventanas_si = []\n",
    "        \n",
    "        img_res = cv2.resize(img[ventana[0][1]:ventana[1][1], ventana[0][0]:ventana[1][0]], (64, 64))\n",
    "        \n",
    "        caracteristicas = obtener_vector_caracteristicas_img(img_res, \n",
    "                                                             orient=9,\n",
    "                                                             pix_per_cell=8,\n",
    "                                                             cell_per_block=2,\n",
    "                                                             hog_channel=3,\n",
    "                                                             color_space=\"RGB\")\n",
    "        \n",
    "        # Estandarizacion de datos\n",
    "        scaler = StandardScaler().fit(caracteristicas)\n",
    "        scaled = scaler.transform(np.array(caracteristicas).reshape(1, -1))\n",
    "        \n",
    "        # Prediccion con SVM\n",
    "        prediccion = model_svc.predict(scaled)\n",
    "        \n",
    "        if prediccion == 1:\n",
    "            ventanas_si.append(ventana)\n",
    "    \n",
    "    return ventanas_si"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def detectar_vehiculos(img):\n",
    "    \n",
    "    ventanas = ventana_deslizante(img,\n",
    "                                  x_ini_fin=[0, 1549],\n",
    "                                  y_ini_fin=[390, 500],\n",
    "                                  xy_ventana=(64, 64), \n",
    "                                  xy_superposicion=(0.5, 0.5))\n",
    "    \n",
    "    ventanas_vehiculos = buscar_vehiculos(img, ventanas)\n",
    "    \n",
    "    return ventanas_vehiculos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_test = mpimg.imread('test0.jpg')\n",
    "ventanas_vehiculos = detectar_vehiculos(img_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "## Extractor de Características: HOP (Histogram Oriented Phase)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/alejandrodebus/anaconda2/envs/cv/lib/python2.7/site-packages/phasepack/tools.py:14: UserWarning: \n",
      "Module 'pyfftw' (FFTW Python bindings) could not be imported. To install it, try\n",
      "running 'pip install pyfftw' from the terminal. Falling back on the slower\n",
      "'fftpack' module for 2D Fourier transforms.\n",
      "  'fftpack' module for 2D Fourier transforms.\"\"\")\n"
     ]
    }
   ],
   "source": [
    "from phasepack import phasecong"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_gray = cv2.cvtColor(img_vehiculo, cv2.COLOR_BGR2GRAY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "M_mod = np.zeros([64,64], dtype=float)\n",
    "step = 64\n",
    "for i in range(0,64,step):\n",
    "    for j in range(0,64,step):\n",
    "        M_mod[i:i+step, j:j+step], _, _, _, _, _, _ = phasecong(img_gray[i:i+step,j:j+step], nscale=2, norient=8, minWaveLength=6, mult=2.1,\n",
    "              sigmaOnf=0.55, k=2., cutOff=0.5, g=10., noiseMethod=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "M, m, img_ori, ft, PC, Eo, T = phasecong(img_gray, nscale=2, norient=8, minWaveLength=6, mult=2.1,\n",
    "              sigmaOnf=0.55, k=2., cutOff=0.5, g=10., noiseMethod=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x198ec8050>"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(M_mod, cmap='gray')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (cv)",
   "language": "python",
   "name": "cv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}