accuracy_assessment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import geopandas as gpd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mplt\n",
    "import numpy as np\n",
    "import pandas as pd \n",
    "import os\n",
    "import shapely as shp\n",
    "import shutil\n",
    "\n",
    "# enable mutiple outputs per notebook cell\n",
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "\n",
    "# set style for plots\n",
    "plt.style.use('classic')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'C:/Users/felix/github/field_boundary_delineation\\\\sample_data\\\\upper_rhine\\\\results\\\\fields.dbf'"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "'C:/Users/felix/github/field_boundary_delineation\\\\sample_data\\\\upper_rhine\\\\results\\\\fields.prj'"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "'C:/Users/felix/github/field_boundary_delineation\\\\sample_data\\\\upper_rhine\\\\results\\\\fields.shp'"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "'C:/Users/felix/github/field_boundary_delineation\\\\sample_data\\\\upper_rhine\\\\results\\\\fields.shx'"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# specify AoI\n",
    "aoi = \"upper_rhine\" #\"magdeburger_boerde\" #\"coastal_landscape\"\n",
    "\n",
    "# specify working dir\n",
    "workdir = \"C:/Users/felix/github/field_boundary_delineation\"\n",
    "\n",
    "# copy delineated field file from ecognition workspace into corresponding sample data folder\n",
    "ecog_dir = os.path.join(\n",
    "    workdir, \n",
    "    \"ecognition_workspace\", \n",
    "    \"results\", \n",
    "    \"ObjectShapes\", \n",
    "    f\"{aoi}.x.subsets.Subset\")\n",
    "\n",
    "sample_dir = os.path.join(workdir, \"sample_data\", aoi)\n",
    "\n",
    "src_files = os.listdir(ecog_dir)\n",
    "for file_name in src_files:\n",
    "    full_file_name = os.path.join(ecog_dir, file_name)\n",
    "    if os.path.isfile(full_file_name):\n",
    "        shutil.copy(full_file_name, os.path.join(sample_dir, \"results\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "# read fields delineated via application as well as manually delineated validation fields\n",
    "fields_dir = os.path.join(sample_dir, \"results\", \"fields.shp\")\n",
    "fields_val_dir = os.path.join(sample_dir, \"manually_delineated_fields\", f\"{aoi}_validation.shp\") \n",
    "\n",
    "fields = gpd.read_file(fields_dir)\n",
    "fields_val = gpd.read_file(fields_val_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Warning: For 1 validation field(s) no overlapping segmented fields were found.\n"
     ]
    }
   ],
   "source": [
    "# spatial join - match delineated and validation fields\n",
    "fields_intersect = fields_val.sjoin(fields, how=\"left\")[[\"geometry\", \"index_right\"]]\n",
    "fields_intersect = fields_intersect.rename(columns = {\"geometry\": \"geom_fields_val\", \"index_right\" : \"idx_fields\"})\n",
    "fields_intersect = pd.merge(fields_intersect, fields, left_on=\"idx_fields\", right_index=True)\n",
    "fields_intersect = fields_intersect.reset_index()\n",
    "fields_intersect = fields_intersect.rename(columns = {\"geometry\": \"geom_fields\", \"index\": \"idx_fields_val\"})\n",
    "fields_intersect = fields_intersect[[\"idx_fields_val\", \"geom_fields_val\", \"idx_fields\", \"geom_fields\"]]\n",
    "\n",
    "if fields_intersect[\"idx_fields_val\"].nunique() < len(fields_val):\n",
    "    missing_fields = len(fields_val)-fields_intersect['idx_fields_val'].nunique()\n",
    "    print(f\"Warning: For {missing_fields} validation field(s) no overlapping segmented fields were found.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>idx_fields_val</th>\n",
       "      <th>geom_fields_val</th>\n",
       "      <th>idx_fields</th>\n",
       "      <th>geom_fields</th>\n",
       "      <th>ov_val</th>\n",
       "      <th>ov_seg</th>\n",
       "      <th>ov_combined</th>\n",
       "      <th>ov_IoU</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>POLYGON ((399670.269 5316421.831, 399879.463 5...</td>\n",
       "      <td>5276.0</td>\n",
       "      <td>POLYGON ((400000.000 5316010.000, 399990.000 5...</td>\n",
       "      <td>69.096641</td>\n",
       "      <td>41.338123</td>\n",
       "      <td>55.217382</td>\n",
       "      <td>34.887905</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>POLYGON ((400569.020 5315977.524, 400559.968 5...</td>\n",
       "      <td>5013.0</td>\n",
       "      <td>POLYGON ((400720.000 5315960.000, 400720.000 5...</td>\n",
       "      <td>66.542722</td>\n",
       "      <td>97.520100</td>\n",
       "      <td>82.031411</td>\n",
       "      <td>65.435452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>POLYGON ((401359.237 5312708.041, 401225.280 5...</td>\n",
       "      <td>7211.0</td>\n",
       "      <td>POLYGON ((401440.000 5312780.000, 401420.000 5...</td>\n",
       "      <td>55.242942</td>\n",
       "      <td>98.262913</td>\n",
       "      <td>76.752927</td>\n",
       "      <td>54.708667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>POLYGON ((401142.827 5317978.821, 401149.067 5...</td>\n",
       "      <td>4097.0</td>\n",
       "      <td>POLYGON ((401450.000 5318060.000, 401440.000 5...</td>\n",
       "      <td>86.632914</td>\n",
       "      <td>28.272834</td>\n",
       "      <td>57.452874</td>\n",
       "      <td>27.091021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>POLYGON ((399853.536 5312241.982, 399924.592 5...</td>\n",
       "      <td>7584.0</td>\n",
       "      <td>POLYGON ((399950.000 5312190.000, 399940.000 5...</td>\n",
       "      <td>86.253642</td>\n",
       "      <td>95.221060</td>\n",
       "      <td>90.737351</td>\n",
       "      <td>82.674748</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   idx_fields_val                                    geom_fields_val  \\\n",
       "0               0  POLYGON ((399670.269 5316421.831, 399879.463 5...   \n",
       "1               1  POLYGON ((400569.020 5315977.524, 400559.968 5...   \n",
       "2               2  POLYGON ((401359.237 5312708.041, 401225.280 5...   \n",
       "3               3  POLYGON ((401142.827 5317978.821, 401149.067 5...   \n",
       "4               4  POLYGON ((399853.536 5312241.982, 399924.592 5...   \n",
       "\n",
       "   idx_fields                                        geom_fields     ov_val  \\\n",
       "0      5276.0  POLYGON ((400000.000 5316010.000, 399990.000 5...  69.096641   \n",
       "1      5013.0  POLYGON ((400720.000 5315960.000, 400720.000 5...  66.542722   \n",
       "2      7211.0  POLYGON ((401440.000 5312780.000, 401420.000 5...  55.242942   \n",
       "3      4097.0  POLYGON ((401450.000 5318060.000, 401440.000 5...  86.632914   \n",
       "4      7584.0  POLYGON ((399950.000 5312190.000, 399940.000 5...  86.253642   \n",
       "\n",
       "      ov_seg  ov_combined     ov_IoU  \n",
       "0  41.338123    55.217382  34.887905  \n",
       "1  97.520100    82.031411  65.435452  \n",
       "2  98.262913    76.752927  54.708667  \n",
       "3  28.272834    57.452874  27.091021  \n",
       "4  95.221060    90.737351  82.674748  "
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate areal overlap measures to find best matching segemented fields\n",
    "# determine overlaps relative to validation fields & segmented fields\n",
    "# additionally intersection over union (aka Jaccard-Index)\n",
    "overlaps_val = []\n",
    "overlaps_seg = []\n",
    "overlaps_IoU = []\n",
    "\n",
    "for index, row in fields_intersect.iterrows():\n",
    "    field_val_geom = shp.geometry.shape(row[\"geom_fields_val\"])\n",
    "    field_geom = shp.geometry.shape(row[\"geom_fields\"])\n",
    "    fields_union = field_val_geom.union(field_geom)\n",
    "    overlaps_val.append((field_val_geom.intersection(field_geom).area/field_val_geom.area)*100)\n",
    "    overlaps_seg.append((field_geom.intersection(field_val_geom).area/field_geom.area)*100)\n",
    "    overlaps_IoU.append((field_geom.intersection(field_val_geom).area/fields_union.area)*100)\n",
    "\n",
    "fields_intersect[\"ov_val\"] = overlaps_val\n",
    "fields_intersect[\"ov_seg\"] = overlaps_seg\n",
    "fields_intersect[\"ov_combined\"] = (fields_intersect[\"ov_seg\"] + fields_intersect[\"ov_val\"])/2\n",
    "fields_intersect[\"ov_IoU\"] = overlaps_IoU\n",
    "\n",
    "# filter delineated fields with maximum overlap for each validation field\n",
    "idx_max_overlap = fields_intersect.groupby(['idx_fields_val'])['ov_combined'].transform(max) == fields_intersect[\"ov_combined\"]\n",
    "fields_intersect = fields_intersect[idx_max_overlap].reset_index(drop=True)\n",
    "\n",
    "fields_intersect.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>idx_fields_val</th>\n",
       "      <th>geom_fields_val</th>\n",
       "      <th>idx_fields</th>\n",
       "      <th>geom_fields</th>\n",
       "      <th>ov_val</th>\n",
       "      <th>ov_seg</th>\n",
       "      <th>ov_combined</th>\n",
       "      <th>ov_IoU</th>\n",
       "      <th>mae_val</th>\n",
       "      <th>mae_seg</th>\n",
       "      <th>mae_combined</th>\n",
       "      <th>hausdorff_dist</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>POLYGON ((399670.269 5316421.831, 399879.463 5...</td>\n",
       "      <td>5276.0</td>\n",
       "      <td>POLYGON ((400000.000 5316010.000, 399990.000 5...</td>\n",
       "      <td>69.096641</td>\n",
       "      <td>41.338123</td>\n",
       "      <td>55.217382</td>\n",
       "      <td>34.887905</td>\n",
       "      <td>56.961182</td>\n",
       "      <td>106.780627</td>\n",
       "      <td>163.741809</td>\n",
       "      <td>268.805133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>POLYGON ((400569.020 5315977.524, 400559.968 5...</td>\n",
       "      <td>5013.0</td>\n",
       "      <td>POLYGON ((400720.000 5315960.000, 400720.000 5...</td>\n",
       "      <td>66.542722</td>\n",
       "      <td>97.520100</td>\n",
       "      <td>82.031411</td>\n",
       "      <td>65.435452</td>\n",
       "      <td>92.917697</td>\n",
       "      <td>10.780272</td>\n",
       "      <td>103.697970</td>\n",
       "      <td>429.087586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>POLYGON ((401359.237 5312708.041, 401225.280 5...</td>\n",
       "      <td>7211.0</td>\n",
       "      <td>POLYGON ((401440.000 5312780.000, 401420.000 5...</td>\n",
       "      <td>55.242942</td>\n",
       "      <td>98.262913</td>\n",
       "      <td>76.752927</td>\n",
       "      <td>54.708667</td>\n",
       "      <td>20.723356</td>\n",
       "      <td>17.442426</td>\n",
       "      <td>38.165782</td>\n",
       "      <td>54.533887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>POLYGON ((401142.827 5317978.821, 401149.067 5...</td>\n",
       "      <td>4097.0</td>\n",
       "      <td>POLYGON ((401450.000 5318060.000, 401440.000 5...</td>\n",
       "      <td>86.632914</td>\n",
       "      <td>28.272834</td>\n",
       "      <td>57.452874</td>\n",
       "      <td>27.091021</td>\n",
       "      <td>3.960631</td>\n",
       "      <td>73.645060</td>\n",
       "      <td>77.605691</td>\n",
       "      <td>199.501035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>POLYGON ((399853.536 5312241.982, 399924.592 5...</td>\n",
       "      <td>7584.0</td>\n",
       "      <td>POLYGON ((399950.000 5312190.000, 399940.000 5...</td>\n",
       "      <td>86.253642</td>\n",
       "      <td>95.221060</td>\n",
       "      <td>90.737351</td>\n",
       "      <td>82.674748</td>\n",
       "      <td>4.528084</td>\n",
       "      <td>5.539923</td>\n",
       "      <td>10.068007</td>\n",
       "      <td>16.759428</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   idx_fields_val                                    geom_fields_val  \\\n",
       "0               0  POLYGON ((399670.269 5316421.831, 399879.463 5...   \n",
       "1               1  POLYGON ((400569.020 5315977.524, 400559.968 5...   \n",
       "2               2  POLYGON ((401359.237 5312708.041, 401225.280 5...   \n",
       "3               3  POLYGON ((401142.827 5317978.821, 401149.067 5...   \n",
       "4               4  POLYGON ((399853.536 5312241.982, 399924.592 5...   \n",
       "\n",
       "   idx_fields                                        geom_fields     ov_val  \\\n",
       "0      5276.0  POLYGON ((400000.000 5316010.000, 399990.000 5...  69.096641   \n",
       "1      5013.0  POLYGON ((400720.000 5315960.000, 400720.000 5...  66.542722   \n",
       "2      7211.0  POLYGON ((401440.000 5312780.000, 401420.000 5...  55.242942   \n",
       "3      4097.0  POLYGON ((401450.000 5318060.000, 401440.000 5...  86.632914   \n",
       "4      7584.0  POLYGON ((399950.000 5312190.000, 399940.000 5...  86.253642   \n",
       "\n",
       "      ov_seg  ov_combined     ov_IoU    mae_val     mae_seg  mae_combined  \\\n",
       "0  41.338123    55.217382  34.887905  56.961182  106.780627    163.741809   \n",
       "1  97.520100    82.031411  65.435452  92.917697   10.780272    103.697970   \n",
       "2  98.262913    76.752927  54.708667  20.723356   17.442426     38.165782   \n",
       "3  28.272834    57.452874  27.091021   3.960631   73.645060     77.605691   \n",
       "4  95.221060    90.737351  82.674748   4.528084    5.539923     10.068007   \n",
       "\n",
       "   hausdorff_dist  \n",
       "0      268.805133  \n",
       "1      429.087586  \n",
       "2       54.533887  \n",
       "3      199.501035  \n",
       "4       16.759428  "
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate additional accuracy measures (edge-based)\n",
    "# define function to calculate mae error\n",
    "def points_along_boundary(geom, dist=2.5):\n",
    "    distance_delta = dist\n",
    "    distances = np.arange(0, geom.length, distance_delta)\n",
    "    points = [geom.exterior.interpolate(distance) for distance in distances]\n",
    "    multipoint = shp.ops.unary_union(points)\n",
    "    return multipoint\n",
    "\n",
    "def mean_absolute_edge_error(source_geom, replica_geom):\n",
    "    dists = []\n",
    "    for ref_point in source_geom.geoms:\n",
    "        near_points = shp.ops.nearest_points(ref_point, replica_geom)\n",
    "        dists.append(near_points[0].distance(near_points[1]))\n",
    "    return np.mean(np.array(dists))\n",
    "\n",
    "# initialise empty lists\n",
    "hausdorff_dists = []\n",
    "mae_val_dists = []\n",
    "mae_seg_dists = []\n",
    "\n",
    "# calculate edge-based metrics for all fields\n",
    "for index, row in fields_intersect.iterrows():\n",
    "    field_val_geom = shp.geometry.shape(row[\"geom_fields_val\"])\n",
    "    field_geom = shp.geometry.shape(row[\"geom_fields\"])\n",
    "    mae_val_dist = mean_absolute_edge_error(\n",
    "        points_along_boundary(field_val_geom),\n",
    "        points_along_boundary(field_geom)\n",
    "        )\n",
    "    mae_seg_dist = mean_absolute_edge_error(\n",
    "        points_along_boundary(field_geom),\n",
    "        points_along_boundary(field_val_geom)\n",
    "        )\n",
    "    mae_val_dists.append(mae_val_dist)\n",
    "    mae_seg_dists.append(mae_seg_dist)\n",
    "    hausdorff_dists.append(field_geom.hausdorff_distance(field_val_geom))\n",
    "\n",
    "fields_intersect[\"mae_val\"] = mae_val_dists\n",
    "fields_intersect[\"mae_seg\"] = mae_seg_dists\n",
    "fields_intersect[\"mae_combined\"] = fields_intersect[\"mae_val\"] + fields_intersect[\"mae_seg\"]\n",
    "fields_intersect[\"hausdorff_dist\"] = hausdorff_dists\n",
    "fields_intersect.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "areal stats\n",
      "ov_val         86.37\n",
      "ov_seg         73.91\n",
      "ov_combined    80.14\n",
      "ov_IoU         64.28\n",
      "dtype: float64\n",
      "\n",
      "edge-based stats\n",
      "mae_val            13.17\n",
      "mae_seg            28.68\n",
      "mae_combined       41.85\n",
      "hausdorff_dist    101.73\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "# summarise stats for all fields\n",
    "print(f\"areal stats\")\n",
    "print(fields_intersect.iloc[:,4:8].mean(numeric_only=True).round(2))\n",
    "\n",
    "print(f\"\\nedge-based stats\")\n",
    "print(fields_intersect.iloc[:,8:].mean(numeric_only=True).round(2))\n",
    "\n",
    "# plot distributions\n",
    "# e.g. for IoU\n",
    "# import seaborn as sns\n",
    "# sns.kdeplot(fields_intersect[\"ov_IoU\"], bw_adjust=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "plotting field idx 4\n",
      "plotting field idx 7\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 560x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 560x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting function - reference & validation fields\n",
    "def plot_fields_acc_metrics(field_intersect_idx=15):\n",
    "    idx = field_intersect_idx\n",
    "    # get stats for field\n",
    "    ov_val = fields_intersect.iloc[idx]['ov_val'].round(1)/100\n",
    "    ov_seg = fields_intersect.iloc[idx]['ov_seg'].round(1)/100\n",
    "    ov_IoU = fields_intersect.iloc[idx]['ov_IoU'].round(1)/100\n",
    "    mae_val = fields_intersect.iloc[idx]['mae_val'].round(1)\n",
    "    mae_seg = fields_intersect.iloc[idx]['mae_seg'].round(1)\n",
    "    mae_combined = fields_intersect.iloc[idx]['mae_combined'].round(1)\n",
    "    acc_stats = (\n",
    "        f\"---------------------------\"\n",
    "        f\"\\naccuracy statistics\"\n",
    "        f\"\\n---------------------------\"\n",
    "        f\"\\n\"\n",
    "        f\"\\narea-based (overlap)\"\n",
    "        f\"\\n * intersection over union: {ov_IoU:.1%}\"\n",
    "        f\"\\n * rel. to validation: {ov_val:.1%}\"\n",
    "        f\"\\n * rel. to extracted: {ov_seg:.1%}\"\n",
    "        f\"\\n\"\n",
    "        f\"\\nedge-based (mae)\"\n",
    "        f\"\\n * combined: {mae_combined:.1f}m\"\n",
    "        f\"\\n * rel. to validation: {mae_val:.1f}m\"\n",
    "        f\"\\n * rel. to extracted: {mae_seg:.1f}m\"\n",
    "        )\n",
    "    # create plot\n",
    "    fig = plt.figure(figsize=(7,5))\n",
    "    gs  = mplt.gridspec.GridSpec(1, 2, width_ratios=[0.25,1])\n",
    "    ax0 = plt.subplot(gs[0])\n",
    "    ax0.axis(\"off\")\n",
    "    ax1 = plt.subplot(gs[1])\n",
    "    xs1, ys1 = fields_intersect[\"geom_fields_val\"].iloc[idx].exterior.xy\n",
    "    xs2, ys2 = fields_intersect[\"geom_fields\"].iloc[idx].exterior.xy\n",
    "    ax1.fill(xs1, ys1, alpha=.75, fc='darkgrey')\n",
    "    ax1.fill(xs2, ys2, alpha=.5, fc='darkblue')\n",
    "    ax1.axis('equal')\n",
    "    ax1.axis(\"off\")\n",
    "    plt.gcf().text(0.01, 0.325, acc_stats, fontsize=10)\n",
    "    return fig;\n",
    "\n",
    "\n",
    "# plot some random examples\n",
    "randints = [np.random.randint(0,len(fields_intersect)) for i in range(2)]\n",
    "\n",
    "for i, idx in enumerate(randints):\n",
    "    print(f\"plotting field idx {idx}\")\n",
    "    p1 = plot_fields_acc_metrics(idx)\n",
    "    #p1.savefig(os.path.join(workdir, \"documentation_figures\", f\"comparison_extr_val_field_{i}.png\"))"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "26de051ba29f2982a8de78e945f0abaf191376122a1563185a90213a26c5da77"
  },
  "kernelspec": {
   "display_name": "Python 3.10.1 64-bit",
   "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": "3.10.1"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}