Fig_04_sim_phaseClosure.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Unwrapping error correction - phase closure - simulation ##\n",
    "+ Figure 4 - Validation and performance evaluation of the phase closure method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Go to directory: /Users/yunjunz/Documents/Paper/2019_MintPy/figs_src/sim_phaseClosure\n",
      "write baseline list info to file: bl_list.txt\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import os\n",
    "import sys\n",
    "import time\n",
    "import pickle\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt, ticker\n",
    "from cvxopt import matrix\n",
    "from mintpy.defaults.plot import *\n",
    "from mintpy.simulation import simulation as sim\n",
    "from mintpy.objects import timeseries, ifgramStack, sensor\n",
    "from mintpy.utils import readfile, ptime, plot as pp, utils as ut, network as pnet\n",
    "from mintpy.utils.solvers import l1regls\n",
    "\n",
    "work_dir = os.path.expanduser('~/Documents/Paper/2019_MintPy/figs_src/sim_phaseClosure')\n",
    "os.chdir(work_dir)\n",
    "print('Go to directory: '+work_dir)\n",
    "\n",
    "# reference file\n",
    "proj_dir = os.path.expanduser('~/data/Galapagos/GalapagosSenDT128/mintpy')\n",
    "sensor_name = sensor.project_name2sensor_name(proj_dir)[0]\n",
    "ts_file = os.path.join(proj_dir, 'timeseries.h5')\n",
    "ts_obj = timeseries(ts_file)\n",
    "ts_obj.open(print_msg=False)\n",
    "ts_obj.save2bl_list_file(out_file='bl_list.txt')\n",
    "\n",
    "# specified time-series displacement\n",
    "tbase = np.array(ts_obj.tbase)\n",
    "ts_sim = sim.sim_variable_timeseries(tbase, display=False)\n",
    "inc_angle = ut.incidence_angle(ts_obj.metadata, dimension=0, print_msg=False)\n",
    "\n",
    "## Decorrelation Noise Setup\n",
    "coh_resid, decor_time = 0.0, 200  # 0.0, 50; 0.1, 50; 0.0, 200; 0.2, 200; 0.4, 200\n",
    "L = 3 * 9"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Network Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "reading existed pickle file: sim_ifgram4phaseClosure.pickle\n",
      "number of acquisitions: 98\n",
      "number of sequential connections: 5\n",
      "number of ifgrams: 475\n",
      "number of triplets: 940\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "re_generate = False\n",
    "num_conn = 5\n",
    "pfile = 'sim_ifgram4phaseClosure.pickle'\n",
    "if not os.path.isfile(pfile) or re_generate:\n",
    "    print('generating ifgram stack with decorrelation noise and unwrapping errors ...')\n",
    "    # simulate ifgram without unwrap error\n",
    "    date12_list = pnet.select_pairs_sequential(ts_obj.dateList, num_connection=num_conn, date12_format='YYYYMMDD_YYYYMMDD')\n",
    "    ifgram_sim = sim.simulate_network(ts_sim, date12_list, decor_time, coh_resid, L=3*9, num_repeat=1, baseline_file='bl_list.txt', sensor_name=sensor_name, inc_angle=inc_angle)[0].flatten()\n",
    "    # add unwrap error\n",
    "    ifgram_err, idx_ifg_err = sim.add_unw_err2ifgram(ifgram_sim, percentage=0.2, Nmax=2, print_msg=True)\n",
    "\n",
    "    # save to pickle file\n",
    "    ds = {}\n",
    "    ds['date12_list'] = date12_list\n",
    "    ds['ifgram_sim'] = ifgram_sim\n",
    "    ds['ifgram_err'] = ifgram_err\n",
    "    ds['idx_ifg_err'] = idx_ifg_err\n",
    "    with open(pfile, 'wb') as f:\n",
    "        pickle.dump(ds, f)\n",
    "    print('saved to pickle file {}.'.format(pfile))\n",
    "    ds = dict(ds)\n",
    "else:\n",
    "    # read pickle file\n",
    "    print('reading existed pickle file:', pfile)\n",
    "    with open(pfile, 'rb') as f:\n",
    "        ds = pickle.load(f)\n",
    "    date12_list = ds['date12_list']\n",
    "    ifgram_sim = ds['ifgram_sim']\n",
    "    ifgram_err = ds['ifgram_err']\n",
    "    idx_ifg_err = ds['idx_ifg_err']\n",
    "\n",
    "num_ifgram = len(date12_list)\n",
    "C = ifgramStack.get_design_matrix4triplet(date12_list=date12_list)\n",
    "print('number of acquisitions: {}'.format(len(ts_obj.dateList)))\n",
    "print('number of sequential connections: {}'.format(num_conn))\n",
    "print('number of ifgrams: {}'.format(num_ifgram))\n",
    "print('number of triplets: {}'.format(C.shape[0]))\n",
    "\n",
    "# plot\n",
    "fig, ax = plt.subplots(figsize=[4, 3])\n",
    "# 1. plot acquisitions\n",
    "ax.plot(ts_obj.times, ts_obj.pbase, 'ko', alpha=0.7, ms=6, mfc=pp.mplColors[1])\n",
    "# 2. plot all interferometric pairs\n",
    "for date12 in date12_list:\n",
    "    date1, date2 = date12.split('_')\n",
    "    idx1 = ts_obj.dateList.index(date1)\n",
    "    idx2 = ts_obj.dateList.index(date2)\n",
    "    x = [ts_obj.times[idx1], ts_obj.times[idx2]]\n",
    "    y = [ts_obj.pbase[idx1], ts_obj.pbase[idx2]]\n",
    "    ax.plot(x, y, '-', lw=2, alpha=0.7, c=pp.mplColors[0])\n",
    "# 3. plot interferometric pairs with error\n",
    "for date12 in np.array(date12_list)[idx_ifg_err]:\n",
    "    date1, date2 = date12.split('_')\n",
    "    idx1 = ts_obj.dateList.index(date1)\n",
    "    idx2 = ts_obj.dateList.index(date2)\n",
    "    x = [ts_obj.times[idx1], ts_obj.times[idx2]]\n",
    "    y = [ts_obj.pbase[idx1], ts_obj.pbase[idx2]]\n",
    "    ax.plot(x, y, '-', lw=2, alpha=1.0, c=pp.mplColors[1])\n",
    "# axis format\n",
    "pp.auto_adjust_xaxis_date(ax, ts_obj.times, fontsize=font_size)\n",
    "ax.set_xlabel('Time [yr]', fontsize=font_size)\n",
    "ax.set_ylabel('Perp baseline [m]', fontsize=font_size)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run unwrapping error correction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error percentage: 0.20 -> 0.02\n"
     ]
    }
   ],
   "source": [
    "ifgram = np.array(ifgram_err)\n",
    "C = ifgramStack.get_design_matrix4triplet(date12_list=date12_list).astype(float)\n",
    "num_tri, num_ifgram = C.shape\n",
    "ifgram = ifgram.reshape(num_ifgram, -1)\n",
    "\n",
    "# 1. calculate the integer ambiguity of the closure phase\n",
    "closure_pha = np.dot(C, ifgram)\n",
    "closure_int = np.round((closure_pha - ut.wrap(closure_pha)) / (2.*np.pi))\n",
    "\n",
    "# 2. solve the integer ambiguity using LASSO regression\n",
    "A = matrix(-C)\n",
    "b = matrix(closure_int)\n",
    "U = np.round(l1regls(A, b, alpha=1e-2, show_progress=0))\n",
    "\n",
    "# 3. correct unwrap error\n",
    "ifgram_cor = ifgram + 2.0*np.pi*U\n",
    "ifgram_cor = ifgram_cor.flatten()\n",
    "\n",
    "# error percentage\n",
    "perc0 = np.sum(np.abs(ifgram_err - ifgram_sim) > 0.1) / num_ifgram\n",
    "perc1 = np.sum(np.abs(ifgram_cor - ifgram_sim) > 0.1) / num_ifgram\n",
    "print('Error percentage: {:.2f} -> {:.2f}'.format(perc0, perc1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# error (orange) / solution (blue) without rounding.\n",
    "x = np.array(l1regls(A, b, alpha=1e-2, show_progress=0)).flatten()\n",
    "fig, ax = plt.subplots(figsize=[18, 5])\n",
    "ax.bar(np.arange(x.size), (ifgram_err - ifgram_sim) / (2.0*np.pi), color=pp.mplColors[1], label='error before')\n",
    "ax.bar(np.arange(x.size), x)\n",
    "ax.yaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "ax.tick_params(which='both', direction='in', labelsize=font_size, bottom=True, top=True, left=True, right=True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fig. 4a - unwrapping-error interferogram number before / after the correction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "saved to file /Users/yunjunz/development/python/2019_RSE_PYSAR/sim_phaseClosure/phaseClosure_err_dist.png\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAACYCAYAAADz/A5OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAFzZJREFUeJzt3XmYXFWdxvHvSxJCSEgwJAbxMYlsBjIMUXABFVwCCMLIA+ISxFGWKMi44O7jSAQFnGEcnJFFNMiuAopLxF1k0VEJIAKaAQJZVHCCQEICCRh+88c5ldyudFfXra6tU+/nee7Tt+5yzrmnq+v0rXPu+SkiMDMzq9cWnS6AmZkNL244zMysFDccZmZWihsOMzMrxQ2HmZmV4obDzMxKccNhZmaluOEwM7NSRtbaKenSOtNZFxEnNKE8ZmbW5VTryXFJ64Az6kjngxExvmmlMjOzrjVYw3FfROw8aCLSooiY0dSSmZlZV6rZcJiZmVVz57iZmZUyWOf4TcCgtyQRsV/TSmRmZl2tZsMBfKWwvhNwLHAJsBSYCvwzcFFrimZmZt2o7j4OSb8GjouIuwvbdgcuioiXtah8ZmbWZco0HCuBKRGxtrBtDPBgRGzbovKZmVmXKdM5fgNwsaRdJI2RtCswH7ipNUUzM7NuVKbheEf+eTewGrgTEPDOJpfJzMy6WOnnOCRtAUwGVkTEMy0plZmZda1GGo5tgEmkuw0AIuL+JpfLzMy61GDDcTfII6iuAPYkPdshNj7jMaL5RTMzs25Upo/jPOB6YCKwCngW8CXSsxzDgqS5nS5DN3F99OX62Mh10Zfro68yDceewEcj4jHSV1wrgQ8Dp7ekZK3R0C9f0mHtOKfdedFAfbTzuho9b3OtjzbXvf9W+vJ7o6BMw7EWGJXXH5Y0NZ+/XSMZDzONVG6jb9B25tWIdl5Xo+dtrvXR7rpvV17DoQ4bMRyuq7EGp8QDgFcB10XExZLOyhmuA5ZFxOGNZN4qkyZNiunTp2+yfcWKFUyePLl0eitXrmTChAktP6fdeTVSH+28rkbP21zro515+W+lr154b9x6660PR0RdF9nQtOp5SO4cYBvg0ohYUzqRvuntQnou5JqIeNtQ0gLYe++9Y+HChUNNxsysZ0i6NSL2rufYukZVSRoB/Aw4KCLW5ec3Lh9CGaudC9zSxPTMzKxF6urjiIj1wPPrPb4MSW8BHiM1TJVtoyWtkrRe0uq8rJe0VtLsZpfBzMzqV6Yh+DRwvqRpkkZI2qKyNJq5pPHAacAHi9sjYh3wPuBnETEuIsYBi4EDI+KnA6Q1V9JCSQtXrFjRf4bzSnwHWObYRo5vVCvz6S/tZtVDu+rH2mM4/D7nTahdzup99VxTt113c8szqfIZmpcBR5LV/QAgG2NzHFPYVnkIsNEHAE8H5kfEcknV+/Yg9XsgaSwpHshdAyUUERcCF0Lq42iwPGZmverhpvZxZM9vsDD9kjQLmA28cIBD9iA9qV5ZfzAiHmlmGczMrLy6G46IWNrkvF8FTAeW5buNccAISbtHxIso3HGQHj68s580zMyszWr2T0iq66lwSZ9uIO8LSV8/zcrLBcD3gYMkTSLNwPuHfOwM4J4G8jAzsyYb7I7j/ZIuojAT7gDeC5xaJuOIeAJ4ovJa0mpgbUSskPQaYHFEPJl3/xn4oKQrI+I3ZfIxM7PmGqzhGAvcx+ANx9pB9g8qIuYV1n8O7Fp4fTZw9lDzMDOzoavZcERE05/bMDOz4c0Ng5mZleKGw8zMSnHDYWZmpbjhMDOzUmp2jtc7D1WeLdfMzHrAYMNx/06ai2ogQ52ryszMhpnBGo6mzk9lZmbD32DPcTR7fqoNJI0GziNNdDiR9KDhJyLiB63K08zMhq7M7LhI+idgf2AShafJI+LtDea9PKe3DDgEuErSHhGxpIH0zMysDeoeVSXpVOBL+ZyjgL8BB5Gi95UWEWsiYl5ELImIZyJiAfAAsJcjAJqZda8yw3GPBQ6IiA8AT+Wfh5GmRh8ySVNI81Pd3bIIgNDciFmVtBpJcyjl6GAUsukf+36fpY9W1G2z022mwSLMNZJeKzQ73VaUs940BzuuWel0UjPK1lgaLYkAuG1EVCLwPSVpVET8VtL+jZSwSNIoUtCmSyJiUd7sCIBmZu1TdwTAMncciyXNzOt3ASdKOgZ4tGzpivKzIpcBTwEnF3YVAzk5AqCZWZcoc8fxSWC7vP5x0h3COOCkRjNXCv03H5gCHBIRTxd2OwKgmVkXKhM69rrC+m+AnZuQ//nAbsDsQtAmHAHQzKx71d1wSNpxoH0RcX/ZjCVNA94FrAMeynHHydsexBEAzcy6Upmvqu4jTS9SjAZY6YQuPeVIfriwVmRBRwA0M+tCZb6q6tORLml7Upzxm5pdKDMz614NT6seEQ8B7wfObF5xzMys2w01HscLgK2bURAzMxseynSO30TfKda3BmYCpzW7UGZm1r3KdI5/per1GuCOiLi3ieUxM7MuV6Zz/JJWFsTMzIaHwULH1vU1VER8qjnFMTOzbjfYHcfzCutbAUcCtwBLganAS4BvtqZoZmbWjQaLAPjOyrqkrwNvjYhvFrYdQYrNYWZmPaLMcNyDgW9XbfsOKXJfy0iaKOlaSWskLZU0p5X5mZlZbWUajvuA91RtO4kUYKmVziVNuT4FOBo4vzC9u5mZtVmZ4bjHA9dK+ghp0sHnAn8HjmhFwWBDAKcjgX+IiNXAzZK+CxwDfKxV+ZqZWQ0RUfcCjAL2A96cf44qc37ZBXgh8GTVtg8B3+vn2LnAQmDh1KlTo6ZTx9fe36ha6Ta6r0VlmPbRBRteT/vogk0OLW7rb/+QytHfcf0d34nfUzOObyStWvXQSLpDKXPZsrTi99SM8g81rVa9/4rp11vWoZalv7yqAEsqn6F5mRsDfDaXueMgUqClG4fcWtVvHLCyattKYJvqA8OhY83MhqLu0LGDPcfxx4jYLa8vp++UIxtExNTSRazPamB81bbxwOMtys/MzAYx2B3HCYX1t7WyIAO4BxgpaZfYOLXJnsDdHSiLmZkx+HMcNxfWb2h9cTbJf42kbwGnSToemAW8Adi33WUxM7Ok7uG4kk6RNCuvv0zSMkn3S9qndcUD0pDfMcD/AV8DTowI33GYmXVImc7xDwDz8/qZwOdJfQ3nAC9tcrk2iIhHgMNblb6ZmZVTpuGYEBErJW1D6meYHRHrJf1Hi8pmZmZdqEzDsVzSvqTgTTfmRmM8sL41RTMzs25UpuH4MHANafqPI/O2Q4HfNrtQZmbWvcoEcroO2KFq89V5MTOzHlHqyXFJuwFvBKZExMnATsCWwO9bUDYzM+tCZYbjHkWabuS5wNvz5nGk0VVmZtYjykyrfhpwQES8m40d4neQRliZmVmPKNNwPJvUUMDGOauCAeavMjOzzVOZhuNWUhyMorfgUVVmZj2lTOf4e4EfSzoOGCvpR8CuwIFlM5U0GjgPmA1MJEUX/ERE/KBsWmZm1l5lhuMukjSD9OzGAmA5sCBSZL5G8l0O7A8sI8Utv0rSHhGxpIH0zMysTepqOCSNIE1xvntEXDXUTCNiDTCvsGmBpAeAvSQ9CKwAxgJP5v1jgKeBQyPip0PN38zMGldXH0dErCeNpBrTikJImkL62uvuiFgHvA/4WUSMi4hxwGLgwFqNhqS5khZKWrhixYraGc6rDirYJK1Kt1kK5Vty1uv7Xe93/1ZzhpRXXcd1c921s2zdVA/dVJYyhmu5odNln1T5DM3L3IEOLNPHcQ7wDUlnAH+iMJoqIu5vtKSSRgFXAJdExKK8eQ/gzrx/LOlBw7tqpePQsWZmQ9Kc0LFVvph/HlC1PYARtU6UdDTwpfzypog4OG/fAriMNP/VyYVT9iA1JpX1B/P06mZm1mFlOsfLDN2tPvcKNjYEAEgSKb7HFOCQiHi6sHvDHQfpAcM7MTOzrlBqrqomOx/YjRTXo9IJjqRJwGTgD3nTDFLHvJmZdYGG7yKGQtI04F2kGOIPSVqdl6OBfwQWFxqTPwNvktSyKINmZla/jtxxRMRSQDUO2bVw7NnA2S0vlJmZ1aUjdxxmZjZ8lWo4JI2S9EpJb86vx+bhsmZm1iPKxOPYg9RJ/WXSaChIU4Zc1IJymZlZlypzx3E+8KmImEGa/gPgBuAVTS+VmZl1rTINx0zg8rwesGHOqZZMQ2JmZt2pTMOxBNiruEHSS0hTopuZWY8oMxz3X4HvS7oA2FLSx4F3Aye0pGRmZtaV6r7jiIgFwMGkp7pvAKYBR0TEj1tUNjMz60KlHgCMiNuAk5pZAEm7kOaiuiYi3tbMtM3MrPnKDMc9RdKsvP4yScsk3S9pnyGW4VzgliGmYWZmbVKmc/wDwAN5/Uzg88BnSXE6GiLpLcBjwM8K20ZLWiVpfWEOq/WS1kqa3WheZmbWHIqoL+aRpFURMV7SNsBSYHJErJf0WERsWzpjaTywEHgtcBywc+WrKknvBN4aEQfm1/cAx0fEjTXSmwvMBZg6depeS5cuLVskM7OeJWkp8HBh04U5QN4myvRxLJe0L+l5jhtzozGeFFK2EacD8yNieQrN0YcjAJqZtVdLIgB+GLiGFK3vyLztUOC3g51YHQEQ+DgwG3jhAKc4AqCZWZcqEwHwOmCHqs1X52Wwc/tEAJT0fmA6sCzfbYwDRkjaPSJehCMAmpl1rbr7ODackPo4JlGIpxER95dMY2tgfGHTh0gNyYmk6Uz+CoyLiCcl/SewRUS8r0T6K0j9MNUm0fc7vHpNAFa24Zx259VIfbTzuho9b3Otj3bm5b+VvnrhvTEtIibXdVZE1LUAuwO3A8+Q+jUqP9fXm0aNtOcBl+f11wD3FPZ9CHgQeGkT8lnY4HkXtuOcDuRVuj7aeV2uj47m5b8VvzcGXMr0cZwHXA+8mjQsdzppWO6vSqTRr4iYV1j/Od0XAfB7bTqn3Xk1op3X1eh5m2t9tLvu25XXcKjDRgyH62rovDLDcR8Fnh0RT1eG4OYRT3dFxPMbybzdJC2MOkcN9ALXR1+uj41cF325Pvoq8wDgWmBUXn9Y0tR8/nZNL1Xr9DsmuYe5PvpyfWzkuujL9VFQ5o7jKuC6iLhY0lnAYcA6YFlEHN7CMpqZWRcpPaoKQNIWwNGkYbSXRgroZGZmPaChhsPMzHpX3aOqJE0kDY2dRbrT2CAi9mtyuczMrEuV6Ry/EtiHNHxrftXS1SRNlHStpDWSlkqa0+kytZKkkyUtlLRO0sVV+14raZGkJyRdL2laYd9oSRfl2YkfknRK2wvfZPma5uff++OSbpd0cGF/T9UHgKTLJT2Yr+seSccX9vVcfUCKC5Rn4L68sG1Oft+skfTt/M9zZV9PfaZsosSDIquA0Y08LNLpBfga8A3SndIrSE9Kzux0uVp4vUcAhwPnAxcXtk/K134UsBXw78CvC/vPJM0l9ixgN+Ah4HWdvp4h1sVY0gOm00n/KB0KPJ5f91x95OuaWflbBmbk69qrV+sjX9uP87VVHkSemd8n++XPjSuBrxeO76nPlE3qq0TF3gzs1OkCN/CGGEuamHHXwrbLgLM6XbY2XPtnqhqOucCvqurmSWBGfv1n4MDC/tOLfyybywL8njRRZ8/XB/AC0swMb+rV+gDeAlxF3xkszgCuLByzU/4c2aaXP1MqS80+DknHFl7+HPihpK+S/tPYICIuqpVOh+1KmhblnsK2O4D9O1SeTppJunYAImKNpMXATEl/JU1ieUfh+DtIdy6bDUlTSO+Ju0lzo/VkfUg6D3gHMIY0ldB1pMBsPVUfOTTEaWyMC1Qxk8KsGBGxWNJTpPfOM/T4Z8pgnePHVL3+E3BA1bYAurnhGMemk3+tJP3n0GvGASuqtlXqYlzhdfW+zYKkUaRZmi+JiEWSerY+IuIkSf9C6rd8FemZrF6sj4HiAtX63FhfY19PqNlwRMSr21WQFlpN35l4ya8f70BZOq1WXawuvF5btW/Yy88eXUb6iuHkvLln6wMgItYDN0t6G+nuq6fqQ9IsBo4LVKsunqmxrycMOqpK0taSzpD0XUnzJI1uR8Ga6B5gpKRdCtv2JH1V0WvuJl070Ce64t0R8Sjpu+49C8dvFvWk9K/kfGAKcGREPJ139WR99GMk+brprfp4FRvjAj1EetzgSEm3sWld7AiMJn2e+DOljo6ji0iBlP6NFL71vzvdMVN2Ab5OGgUxFng5m/kICNIHwVakUTCX5fWRwOR87UfmbZ+j76iZs4AbSKNmZpA+KIb9qBngAuDXpBgvxe09Vx/As0mdweOAEcBBwBrgDb1WH8DWwPaF5WxSlNPJpD6OVcAr8+fG5fQdVdVTnymb1F0dlfsg8Jy8/jzggU4XuoE3yETg2/kPZBkwp9NlavH1ziP1PRWXeXnfbGARabTML4DphfNG538UVpGCaZ3S6WtpQl1My9e/lvT1Q2U5ukfrY3L+8H8sX9edwAmF/T1VH1V1M488qiq/npM/L9YA3wEmFvb11GdK9TLolCOSVkXE+MLrRyJiYq1zzMxs81XPlCMjJb2ajaFiq18TKfiSmZn1gHruOJaQbvUHEhGxYzMLZWZm3cuz45qZWSllJjk0MzNzw2FmZuW44TAzs1LccJjVIGmMpO9JWinp6k6Xx6wbuOGwYUXSEkmz6zz2F8UgRQ16I2mqku0i4qghpjWsSPpLbjhfI+lbnS6PdQ83HGYDkDSC9OT5PRHx9wbOrzs0cyfSGySv5wEPR8STpCBPt7Urb+t+bjhs2JL0Dkk3Szpb0qOSHqiEhZX0WdI8Q1+UtFrSF/P2GZJ+IukRSf8r6U2F9C6WdL6k6yStAW4EPgW8OadxXD7uWEl/zHn+qCq8akh6j6R7gXvztn0l3ZK/7rpF0r6F458v6UalsLY/lXRuJXyppOk5veMkLSPFxEHS1Tl068p87syqazhP0g9ymX8paXtJ5+TyLpLU32yw1fYGbi2su+GwjTo954kXL2UWYAkwO6+/A3gaOIE0Yd+JwF/Y+HzSL4DjC+eOBZYD7yTNmvAi4GHy5HTAxaTJ6l5O+qdqKzadv+hw4D5S6NSRwCfpGzUvgJ+Q5jIak38+SoptMxJ4a369XT7+f0iT621JCkG6io1R6Kbn9C7NZR+Ttx9Liv0wGjgH+F0h/4vzNe2Vy/9z4AHg7bmOPgNcX6N+TyXNY7UWeCKvV+JPPAaM6PR7wEvnF99x2HC3NCK+HCm2xCXAc0h9Ev05FFgSEV+NiL9HxG3AN0n9GBXfiYhfRsQzEbG2nzTeBZwZEX+M9PXVGcCs4l1H3v9IpK95Xg/cGxGX5Ty/RppE8DBJU4EXA5+KiKci4mbgu/3kOS8i1uT0iIiLIuLxiFhHatj2lDShcPy1EXFrLv+1wNqIuDTX0TfoP/4EOe1Pk2KPPwDsCBwC/DAiJkTEtjkN63FuOGy42xDGOCKeyKvjBjh2GvBSSY9VFuBo0pTaFcsHyW8a8IXC+Y+Q5m177gBp7AAsrUpjaT5+B+CRQrkHyn/DNkkjJJ0labGkVaQ7MEgf9hV/Law/2c/rfutH0qx8TY8CO5MauOuBV+XrPaK/86z3tK2zzawDqufTWQ7cEBHV4Y9rnVNtOfDZiLiizjT+QmpsiqYCPySFLJgoaetC4/G8QdKbQ4qdMZvUaEwgfdBr09PKiYjfAdtK+hjp674zJd0OHBUR9w01fdt8+I7DNmd/JX3dUrEA2FXSMZJG5eXFknYrkeYFwMcrHdKSJkiqNUz3upznHEkjJb0Z2B1YEBFLgYXAPElbStoHOGyQ/LchxQf/GykQ0Rklyl6vvYDbJG1JujNa3II8bBhzw2Gbsy8Ab8yjif4rIh4HDiRFwPsL6Wuuz5E6mesSEdfmc76evyq6Czi4xvF/I/WtfJD0Yf8R4NCIeDgfcjSwT973GVIfxLoaRbiU9FXXn4E/kCIbNltl+O0ewJ0R4ZlQrQ/PjmvWRSR9A1gUEad2uixmA/Edh1kH5a/KdpK0haTXkfovvt3pcpnV4s5xs87aHvgWsB3wJ+DEiLi9s0Uyq81fVZmZWSn+qsrMzEpxw2FmZqW44TAzs1LccJiZWSluOMzMrJT/B5OZndOVz3SpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x108 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=[6, 1.5])\n",
    "wid = 1\n",
    "x = np.arange(num_ifgram)+1 - wid/2.\n",
    "ax.bar(x, ifgram_err - ifgram_sim, width=wid, color=pp.mplColors[1])\n",
    "ax.bar(x, ifgram_cor - ifgram_sim, width=wid, color=pp.mplColors[0])\n",
    "\n",
    "# axis format\n",
    "#ax.xaxis.tick_top()\n",
    "#ax.xaxis.set_label_position('top')\n",
    "ax.tick_params(which='both', direction='out', labelsize=font_size, bottom=True, top=True, left=True, right=True)\n",
    "ax.xaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "ax.set_yticks([-4*np.pi, -2*np.pi, 0, 2*np.pi, 4*np.pi])\n",
    "ax.set_yticklabels([r'-4$\\pi$', r'-2$\\pi$', '0', r'2$\\pi$', r'4$\\pi$'])\n",
    "ax.set_xlabel('Interferogram #', fontsize=font_size)\n",
    "ax.set_ylabel('Phase residual [rad]', fontsize=font_size)\n",
    "ax.set_xlim(-10, ifgram_sim.shape[0]+10)\n",
    "\n",
    "# output\n",
    "out_file = os.path.join(work_dir, 'phaseClosure_err_dist.png')\n",
    "plt.savefig(out_file, bbox_inches='tight', transparent=True, dpi=fig_dpi)\n",
    "print('saved to file', out_file)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fig. 4b - performance as a function of input error percentage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stats_02cycle_03conn.pickle\n",
      "stats_02cycle_05conn.pickle\n",
      "stats_02cycle_10conn.pickle\n",
      "saved to file /Users/yunjunz/development/python/2019_RSE_PYSAR/sim_phaseClosure/phaseClosure_performance_redun.png\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=[3, 2])\n",
    "\n",
    "# plot various number of connections\n",
    "num_cycle = '02'\n",
    "for i, num_conn in enumerate(['03', '05', '10']):\n",
    "    pfile = 'stats_{}cycle_{}conn.pickle'.format(num_cycle, num_conn)\n",
    "    print(pfile)\n",
    "    with open(pfile, 'rb') as f:\n",
    "        ds = pickle.load(f)\n",
    "    ax.plot(ds['percentage0'], ds['percentage1'], 'o', ms=2**2, color=pp.mplColors[i], label=int(num_conn))\n",
    "    \n",
    "ax.plot(np.array([0,1]), np.array([0,1]), 'k--')\n",
    "\n",
    "# axis format\n",
    "ax.set_xlim(0,1.0)\n",
    "ax.set_ylim(0,0.5)\n",
    "ax.set_xlabel('Input percentage [%]', fontsize=font_size)\n",
    "ax.set_ylabel('Output percentage [%]', fontsize=font_size)\n",
    "#ax.yaxis.tick_right()\n",
    "#ax.yaxis.set_label_position(\"right\")\n",
    "ax.tick_params(which='both', direction='out', labelsize=font_size, bottom=True, top=True, left=True, right=True)\n",
    "ax.xaxis.set_major_locator(ticker.MaxNLocator(6))\n",
    "ax.xaxis.set_minor_locator(ticker.AutoMinorLocator(4))\n",
    "ax.yaxis.set_major_locator(ticker.MaxNLocator(3))\n",
    "ax.yaxis.set_minor_locator(ticker.AutoMinorLocator(4))\n",
    "ax.set_xticklabels(['{:d}'.format(int(x*100.)) for x in ax.get_xticks()])\n",
    "ax.set_yticklabels(['{:d}'.format(int(x*100.)) for x in ax.get_yticks()])\n",
    "#ax.legend(loc='best', frameon=False, fontsize=font_size, borderaxespad=-0.5, handletextpad=-0.3)\n",
    "ax.legend(loc=(0.72,0.15), ncol=1, columnspacing=0.3, frameon=False, fontsize=font_size, handletextpad=-0.3)\n",
    "\n",
    "# save\n",
    "out_file = os.path.join(work_dir, 'phaseClosure_performance_redun.png')\n",
    "plt.savefig(out_file, bbox_inches='tight', transparent=True, dpi=fig_dpi)\n",
    "print('saved to file', out_file)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fig. 4c - performance as a function of input error magnitude"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stats_02cycle_05conn.pickle\n",
      "stats_05cycle_05conn.pickle\n",
      "stats_10cycle_05conn.pickle\n",
      "saved to file /Users/yunjunz/development/python/2019_RSE_PYSAR/sim_phaseClosure/phaseClosure_performance_maxMag.png\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=[3, 2])\n",
    "\n",
    "# plot various number of connections\n",
    "num_conn = '05'\n",
    "for i, num_cycle in enumerate(['02', '05', '10']):\n",
    "    pfile = 'stats_{}cycle_{}conn.pickle'.format(num_cycle, num_conn)\n",
    "    print(pfile)\n",
    "    with open(pfile, 'rb') as f:\n",
    "        ds = pickle.load(f)\n",
    "    ax.plot(ds['percentage0'], ds['percentage1'], 'o', ms=2**2, color=pp.mplColors[i], label=int(num_cycle))\n",
    "    \n",
    "ax.plot(np.array([0,1]), np.array([0,1]), 'k--')\n",
    "\n",
    "# axis format\n",
    "ax.set_xlim(0,1)\n",
    "ax.set_ylim(0,0.5)\n",
    "ax.set_xlabel('Input percentage [%]', fontsize=font_size)\n",
    "ax.set_ylabel('Output percentage [%]', fontsize=font_size)\n",
    "ax.yaxis.tick_right()\n",
    "ax.yaxis.set_label_position(\"right\")\n",
    "ax.tick_params(which='both', direction='out', labelsize=font_size, bottom=True, top=True, left=True, right=True)\n",
    "ax.xaxis.set_major_locator(ticker.MaxNLocator(6))\n",
    "ax.xaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "ax.set_xticklabels(['{:d}'.format(int(x*100.)) for x in ax.get_xticks()])\n",
    "ax.set_yticklabels(['{:d}'.format(int(x*100.)) for x in ax.get_yticks()])\n",
    "ax.legend(loc=(0.72,0.15), frameon=False, fontsize=font_size, handletextpad=-0.3)\n",
    "\n",
    "# save\n",
    "out_file = os.path.join(work_dir, 'phaseClosure_performance_maxMag.png')\n",
    "plt.savefig(out_file, bbox_inches='tight', transparent=True, dpi=fig_dpi)\n",
    "print('saved to file', out_file)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input /      output error:\n",
      "        3conn   5conn   10conn\n",
      "0.00\t0.000\t0.000\t0.000\n",
      "0.05\t0.002\t0.000\t0.000\n",
      "0.10\t0.012\t0.000\t0.000\n",
      "0.15\t0.037\t0.003\t0.000\n",
      "0.20\t0.068\t0.007\t0.001\n",
      "0.25\t0.120\t0.017\t0.001\n",
      "0.30\t0.165\t0.040\t0.002\n",
      "0.35\t0.233\t0.076\t0.004\n",
      "0.40\t0.287\t0.125\t0.013\n",
      "0.45\t0.345\t0.183\t0.024\n",
      "0.50\t0.405\t0.256\t0.055\n",
      "0.55\t0.456\t0.335\t0.092\n",
      "0.60\t0.505\t0.405\t0.176\n",
      "0.65\t0.551\t0.455\t0.274\n",
      "0.70\t0.585\t0.519\t0.375\n",
      "0.75\t0.617\t0.562\t0.448\n",
      "0.80\t0.648\t0.598\t0.501\n",
      "0.85\t0.673\t0.631\t0.547\n",
      "0.90\t0.705\t0.656\t0.582\n",
      "0.95\t0.723\t0.683\t0.615\n",
      "1.00\t0.746\t0.695\t0.637\n"
     ]
    }
   ],
   "source": [
    "## print\n",
    "ds3 = pickle.load(open('stats_02cycle_03conn.pickle', 'rb'))\n",
    "ds5 = pickle.load(open('stats_02cycle_05conn.pickle', 'rb'))\n",
    "ds10 = pickle.load(open('stats_02cycle_10conn.pickle', 'rb'))\n",
    "print('input /      output error:')\n",
    "print('        3conn   5conn   10conn')\n",
    "for i in range(len(ds5['percentage0'])):\n",
    "    print('{:.2f}\\t{:.3f}\\t{:.3f}\\t{:.3f}'.format(ds5['percentage0'][i], ds3['percentage1'][i], ds5['percentage1'][i], ds10['percentage1'][i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done.\n"
     ]
    }
   ],
   "source": [
    "def correct_unwrap_error_closure(ifgram, C):\n",
    "    \"\"\"Estimate unwrapping error from a stack of unwrapped interferometric phase\n",
    "    Parameters: ifgram : 2D np.array in size of (num_ifgram, num_pixel) of unwrap phase in float32\n",
    "                C      : 2D np.array in size of (num_triangle, num_ifgram) triangle design matrix in float64\n",
    "    Returns:    ifgram_cor : 2D np.array in size of (num_ifgram, num_pixel) of unwrap phase after correction\n",
    "                U          : 2D np.array in size of (num_ifgram, num_pixel) of phase jump integer\n",
    "    Example:    ifgram_cor, U = estimate_unwrap_error(ifgram, C)\n",
    "    \"\"\"\n",
    "    ifgram = ifgram.reshape(C.shape[1], -1)\n",
    "\n",
    "    # 1. calculate the integer ambiguity of the closure phase\n",
    "    closure_pha = np.dot(C, ifgram)\n",
    "    closure_int = np.round((closure_pha - ut.wrap(closure_pha)) / (2.*np.pi))\n",
    "\n",
    "    # 2. solve the integer ambiguity using LASSO regression\n",
    "    A = matrix(-C)\n",
    "    b = matrix(closure_int)\n",
    "    U = np.round(l1regls(A, b, alpha=1e-2, show_progress=0))\n",
    "\n",
    "    # 3. correct unwrap error\n",
    "    ifgram_cor = ifgram + 2.0*np.pi*U\n",
    "    return ifgram_cor, U\n",
    "\n",
    "\n",
    "def phase_closure_performance(num_conn=3, Nmax=2, step=0.05, num_rep=100):\n",
    "    # simulate ifgram without unw err\n",
    "    date12_list = pnet.select_pairs_sequential(ts_obj.dateList, num_connection=num_conn, date12_format='YYYYMMDD_YYYYMMDD')\n",
    "    num_ifgram = len(date12_list)\n",
    "    C = ifgramStack.get_design_matrix4triplet(date12_list=date12_list).astype(float)\n",
    "    #print('number of connections:', num_conn)\n",
    "    print('number of ifgrams: {}'.format(num_ifgram))\n",
    "    print('number of triplets: {}'.format(C.shape[0]))\n",
    "\n",
    "    ifgram = sim.simulate_network(ts_sim, date12_list, decor_time, coh_resid,\n",
    "                                  L=L, num_repeat=1,\n",
    "                                  baseline_file='bl_list.txt',\n",
    "                                  sensor_name=sensor_name,\n",
    "                                  inc_angle=inc_angle)[0].flatten()\n",
    "\n",
    "    # calculate percentage change for various percentage of unw err\n",
    "    print('calculating the improvement of unw err correction for various percentage of errors ...')\n",
    "    percentage0s = np.arange(0, 1+step, step)\n",
    "    num_step = percentage0s.shape[0]\n",
    "    percentage1s = np.zeros(num_step, np.float32)\n",
    "    prog_bar = ptime.progressBar(maxValue=num_step)\n",
    "    for i in range(num_step):\n",
    "        perc1 = 0\n",
    "        for j in range(num_rep):\n",
    "            ifgram_err = sim.add_unw_err2ifgram(ifgram, percentage=percentage0s[i], Nmax=Nmax, print_msg=False)[0]\n",
    "            ifgram_cor = correct_unwrap_error_closure(ifgram_err, C)[0].flatten()\n",
    "            perc1 += np.sum(np.abs(ifgram_cor - ifgram) > 0.1) / num_ifgram\n",
    "        perc1 /= num_rep\n",
    "        percentage1s[i] = perc1\n",
    "        prog_bar.update(i+1)\n",
    "    prog_bar.close()\n",
    "    return percentage1s\n",
    "\n",
    "print('Done.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of ifgrams: 288\n",
      "number of triplets: 286\n",
      "calculating the improvement of unw err correction for various percentage of errors ...\n",
      "[==================================================]  124s /     2s \n",
      "Done.\n"
     ]
    }
   ],
   "source": [
    "## Run performance\n",
    "test = True\n",
    "\n",
    "step = 0.05\n",
    "Nmax = 2 # 2, 5, 10\n",
    "percentage0s = np.arange(0, 1+step, step)\n",
    "for num_conn in [3]:   #[3,5,10]\n",
    "    out_file = 'stats_{:02d}cycle_{:02d}conn.pickle'.format(Nmax, num_conn)\n",
    "    if test or not os.path.isfile(out_file):\n",
    "        percentage1s = phase_closure_performance(num_conn, Nmax, step=step, num_rep=100)\n",
    "\n",
    "        if not test:\n",
    "            # save to pickle file\n",
    "            ds = {}\n",
    "            ds['num_conn'] = num_conn\n",
    "            ds['percentage0'] = percentage0s\n",
    "            ds['percentage1'] = percentage1s\n",
    "            with open(out_file, 'wb') as f:\n",
    "                pickle.dump(ds, f)\n",
    "            print('saved to pickle file {}.'.format(out_file))\n",
    "            ds = dict(ds)\n",
    "    else:\n",
    "        print('{} exists, skip to the next.'.format(out_file))\n",
    "print('Done.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# simple plot for performance result\n",
    "\n",
    "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=[4, 3])\n",
    "ax.plot(percentage0s, percentage1s, 'o', ms=2**2, color=pp.mplColors[0], label='percentange 1')\n",
    "\n",
    "ax.plot(np.array([0,1]), np.array([0,1]), 'k--')\n",
    "ax.tick_params(which='both', direction='in', labelsize=font_size, bottom=True, top=True, left=True, right=True)\n",
    "ax.set_xlim(0,1)\n",
    "ax.set_ylim(0,1)\n",
    "ax.xaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(ticker.AutoMinorLocator())\n",
    "plt.legend()\n",
    "# save\n",
    "#out_file = 'plot_stats_{:02d}cycle_D{:d}.png'.format(Nmax, int(D))\n",
    "#plt.savefig(out_file, bbox_inches='tight', transparent=True, dpi=fig_dpi)\n",
    "#print('saved to file', out_file)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}