{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Exercice 2 : Importation d'images satellitaires\n",
"\n",
"Ce notebook illustre l'utilisation de données satellite réelles Sentinel-2 depuis le **Microsoft Planetary Computer**.\n",
" \n",
"**Source** : Microsoft Planetary Computer (STAC API) \n",
"**Librairies** : rasterio, pandas, matplotlib, seaborn, numpy, planetary-computer, pystac-client\n",
"\n",
"**Explorer le catalogue** : https://planetarycomputer.microsoft.com/explore"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [],
"source": [
"import rasterio\n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"import seaborn as sns\n",
"import numpy as np\n",
"import planetary_computer\n",
"from pystac_client import Client"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 1. Connexion à Planetary Computer\n",
"\n",
"Accès au catalogue STAC pour télécharger une image Sentinel-2."
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Connexion au Planetary Computer réussie\n",
"Item Sentinel-2 trouvé: S2B_MSIL2A_20250824T185919_R013_T11UMT_20250824T224700\n",
"4 bandes spectrales récupérées\n"
]
}
],
"source": [
"# Connexion au Planetary Computer\n",
"catalog = Client.open(\n",
" \"https://planetarycomputer.microsoft.com/api/stac/v1\",\n",
" modifier=planetary_computer.sign_inplace\n",
")\n",
"print(\"Connexion au Planetary Computer réussie\")\n",
"\n",
"# Récupérer un item Sentinel-2 spécifique\n",
"item = catalog.get_collection(\"sentinel-2-l2a\").get_item(\n",
" \"S2B_MSIL2A_20250824T185919_R013_T11UMT_20250824T224700\"\n",
")\n",
"print(f\"Item Sentinel-2 trouvé: {item.id}\")\n",
"\n",
"# Extraire les URLs des bandes spectrales\n",
"# B02=Bleu, B03=Vert, B04=Rouge, B08=NIR (Near Infrared)\n",
"bands = {k: item.assets[k].href for k in ['B02', 'B03', 'B04', 'B08']}\n",
"print(f\"4 bandes spectrales récupérées\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 2. Chargement des bandes\n",
"\n",
"Lecture des bandes spectrales (bleu, vert, rouge, NIR) avec masquage des valeurs invalides."
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Chargement: peut prendre 30-60 secondes)\n",
"Bande bleue chargée (B02)\n",
"Bande verte chargée (B03)\n",
"Bande rouge chargée (B04)\n",
"Bande NIR chargée (B08)\n",
"\n",
"Toutes les bandes chargées et masquées\n",
" Dimensions: 10980 × 10980 pixels\n",
" Pixels valides: 118,331,336 / 120,560,400 (98.2%)\n"
]
}
],
"source": [
"\n",
"print(\"Chargement: peut prendre 30-60 secondes)\")\n",
"# Charger la bande bleue (B02) et extraire les métadonnées\n",
"with rasterio.open(bands['B02']) as src:\n",
" blue = src.read(1) # Lire la première (et unique) bande\n",
" transform = src.transform # Transformation affine (géoréférencement)\n",
" crs = src.crs # Système de coordonnées (ex: EPSG:32618 pour UTM Zone 18N)\n",
" nodata = src.nodata # Valeur représentant pixels invalides (nuages, ombres)\n",
"print(f\"Bande bleue chargée (B02)\")\n",
"\n",
"# Charger les autres bandes (réutilise les métadonnées de B02)\n",
"with rasterio.open(bands['B03']) as src:\n",
" green = src.read(1)\n",
"print(f\"Bande verte chargée (B03)\")\n",
"with rasterio.open(bands['B04']) as src:\n",
" red = src.read(1)\n",
"print(f\"Bande rouge chargée (B04)\")\n",
"with rasterio.open(bands['B08']) as src:\n",
" nir = src.read(1)\n",
"print(f\"Bande NIR chargée (B08)\")\n",
"\n",
"# Masquer les valeurs NoData \n",
"mask = (blue != nodata) & (green != nodata) & (red != nodata) & (nir != nodata)\n",
"\n",
"# Remplacer pixels invalides par NaN pour calculer les stats\n",
"blue_masked = np.where(mask, blue, np.nan)\n",
"green_masked = np.where(mask, green, np.nan)\n",
"red_masked = np.where(mask, red, np.nan)\n",
"nir_masked = np.where(mask, nir, np.nan)\n",
"\n",
"print(f\"\\nToutes les bandes chargées et masquées\")\n",
"print(f\" Dimensions: {red.shape[0]} × {red.shape[1]} pixels\")\n",
"print(f\" Pixels valides: {np.sum(mask):,} / {mask.size:,} ({100*np.sum(mask)/mask.size:.1f}%)\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 3. Statistiques avec la librarie Pandas"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"\n",
"\n",
"
\n",
" \n",
" \n",
" | \n",
" Bande | \n",
" Min | \n",
" Max | \n",
" Moyenne | \n",
"
\n",
" \n",
" \n",
" \n",
" | 0 | \n",
" Bleu | \n",
" 1.0 | \n",
" 18144.0 | \n",
" 1960.8 | \n",
"
\n",
" \n",
" | 1 | \n",
" Vert | \n",
" 1.0 | \n",
" 17472.0 | \n",
" 2121.8 | \n",
"
\n",
" \n",
" | 2 | \n",
" Rouge | \n",
" 203.0 | \n",
" 17008.0 | \n",
" 2080.1 | \n",
"
\n",
" \n",
" | 3 | \n",
" NIR | \n",
" 1.0 | \n",
" 16544.0 | \n",
" 3114.8 | \n",
"
\n",
" \n",
"
\n",
"