LilacSat-1 image repair.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "PATH = 'lilacsat1-data/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "yy = np.fromfile(PATH + '654_dk3wn.jpg', dtype='uint8')[:0x3b72]\n",
    "y = np.fromfile(PATH + '654.jpg', dtype='uint8')[:0x3b72]\n",
    "x = np.fromfile(PATH + '654_dk3wn_2.jpg', dtype='uint8')[:0x3b72]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7edab5908>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAdOOllM12uIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((x==0)[:x.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7ed9cf400>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((yy==0)[:yy.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "missing = np.all((x==0)[:x.size//64*64].reshape((-1,64)), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7ed9aceb8>"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((y==0)[:y.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "z = x.copy()\n",
    "z[:z.size//64*64].reshape((-1,64))[missing,:] = y[:y.size//64*64].reshape((-1,64))[missing,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "missing = np.all((z==0)[:z.size//64*64].reshape((-1,64)), axis=1)\n",
    "z[:z.size//64*64].reshape((-1,64))[missing,:] = yy[:yy.size//64*64].reshape((-1,64))[missing,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7ed90e080>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVEAAAD8CAYAAADOg5fGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((z==0)[:z.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "z.tofile(PATH + '654_fixed.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x = np.fromfile(PATH + '653.jpg', dtype='uint8')[:0x4324]\n",
    "y = np.fromfile(PATH + '653_dk3wn.jpg', dtype='uint8')[:0x4324]\n",
    "yy = np.fromfile(PATH + '653_2.jpg', dtype='uint8')[:0x4324]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eee44128>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAS8AAAD8CAYAAAAmAyLtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((x==0)[:x.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "missing = np.all((x==0)[:x.size//64*64].reshape((-1,64)), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eee1e668>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD6CAYAAACvZ4z8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4Uz0klQ4E70kFc5EL0mF+3/Iw1cjiffw+wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((y==0)[:y.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib64/python3.6/site-packages/ipykernel_launcher.py:2: VisibleDeprecationWarning: boolean index did not match indexed array along dimension 0; dimension is 206 but corresponding boolean dimension is 268\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "z = x.copy()\n",
    "z[:z.size//64*64].reshape((-1,64))[missing,:] = y[:y.size//64*64].reshape((-1,64))[missing,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eed81a58>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAS8AAAD8CAYAAAAmAyLtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUpeIpIkJS8RSZKSl4gkSclLRJL0/5HZGDIXO6rlAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((z==0)[:z.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eed65160>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAS8AAAD8CAYAAAAmAyLtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKbxEJEn/D4hfc1lSYFY2AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((yy==0)[:yy.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "missing = np.all((z==0)[:z.size//64*64].reshape((-1,64)), axis=1)\n",
    "z[:z.size//64*64].reshape((-1,64))[missing,:] = yy[:yy.size//64*64].reshape((-1,64))[missing,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eecc0518>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAS8AAAD8CAYAAAAmAyLtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJSJJUniJSJIUXiKSpP8H4IoKpmXbroIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((z==0)[:z.size//64*64].reshape((-1,64)), aspect = 0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eec96b00>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAADZCAYAAAAwlsvXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJDXCQJekRhjoktQIA12SGmGgS1Ij/h+PiV7sVvHGMQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((z!=yy)[:z.size//64*64].reshape((-1,64))[105:140,:], aspect = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "z[:z.size//64*64].reshape((-1,64))[105:107,:] = y[:y.size//64*64].reshape((-1,64))[105:107,:]\n",
    "z[:z.size//64*64].reshape((-1,64))[129,:] = yy[:y.size//64*64].reshape((-1,64))[129,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fa7eebfa518>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAABCCAYAAAB3uzx2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nPAWJwwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow((z!=yy)[:z.size//64*64].reshape((-1,64))[129:135,:], aspect = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "z.tofile(PATH + '653_fixed.jpg')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python3",
   "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
}