Upload 2 files
Browse files
api.py
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from sentinelsat import SentinelAPI, read_geojson, geojson_to_wkt
|
2 |
+
from datetime import datetime, timedelta,date
|
3 |
+
import zipfile
|
4 |
+
import rasterio
|
5 |
+
from rasterio.plot import show
|
6 |
+
from PIL import Image
|
7 |
+
import matplotlib.pyplot as plt
|
8 |
+
import numpy as np
|
9 |
+
import pandas as pd
|
10 |
+
import os
|
11 |
+
from glob import glob
|
12 |
+
from tqdm import tqdm
|
13 |
+
#from haversine import haversine, Unit
|
14 |
+
#from xml.etree import ElementTree as et
|
15 |
+
import xmltodict
|
16 |
+
import json
|
17 |
+
import warnings
|
18 |
+
import shutil
|
19 |
+
from shapely.geometry import Point
|
20 |
+
from shapely.geometry.polygon import Polygon
|
21 |
+
warnings.filterwarnings('ignore')
|
22 |
+
|
23 |
+
##
|
24 |
+
def unzip():
|
25 |
+
files = glob('*.zip')
|
26 |
+
for file in files:
|
27 |
+
with zipfile.ZipFile(file, 'r') as zip_ref:
|
28 |
+
zip_ref.extractall()
|
29 |
+
|
30 |
+
|
31 |
+
##
|
32 |
+
def select_best_cloud_coverage_tile():
|
33 |
+
tile_names = {}
|
34 |
+
cld_prob = []
|
35 |
+
folders = glob('*.SAFE')
|
36 |
+
for fold in folders:
|
37 |
+
metadata_path = fold+"/MTD_MSIL2A.xml"
|
38 |
+
xml_file=open(metadata_path,"r")
|
39 |
+
xml_string=xml_file.read()
|
40 |
+
python_dict=xmltodict.parse(xml_string)
|
41 |
+
cld = float(python_dict["n1:Level-2A_User_Product"]["n1:Quality_Indicators_Info"]["Cloud_Coverage_Assessment"])
|
42 |
+
tile_names[cld] = fold
|
43 |
+
cld_prob.append(cld)
|
44 |
+
name = tile_names[min(cld_prob)]
|
45 |
+
dates = name.split('_')[2][:8]
|
46 |
+
acquisition_date = datetime.strptime(dates, "%Y%m%d")
|
47 |
+
today = datetime.now()
|
48 |
+
delta = (today - acquisition_date)
|
49 |
+
days_ago = delta.days
|
50 |
+
return name,min(cld_prob),days_ago
|
51 |
+
|
52 |
+
##
|
53 |
+
def find_good_tile(df, point):
|
54 |
+
for row in tqdm(df.itertuples(),total=len(df)):
|
55 |
+
tile_name = row.tiles
|
56 |
+
coordinate = row.coords
|
57 |
+
x = coordinate.replace(']','')
|
58 |
+
x = x.replace('[','')
|
59 |
+
x = x.replace("'",'')
|
60 |
+
tab = [[float(x.split(",")[i]),float(x.split(",")[i+1])] for i in range(0,len(x.split(",")),2)]
|
61 |
+
polygon = Polygon(tab)
|
62 |
+
result = polygon.contains(point)
|
63 |
+
if result:
|
64 |
+
print(tile_name)
|
65 |
+
return "S2B_MSIL2A_20230104T103329_N0509_R108_T30NUL_20230104T132032.SAFE.zip" #tile_name
|
66 |
+
|
67 |
+
return "S2B_MSIL2A_20230104T103329_N0509_R108_T30NUL_20230104T132032.SAFE.zip" #404
|
68 |
+
|
69 |
+
##
|
70 |
+
def delete_tiles():
|
71 |
+
files = glob('*.zip')
|
72 |
+
folders1 = glob('*.SAFE')
|
73 |
+
folders2 = glob("*.tmp")
|
74 |
+
for f in files:
|
75 |
+
os.remove(f)
|
76 |
+
for fold in folders1:
|
77 |
+
shutil.rmtree(fold, ignore_errors=True)
|
78 |
+
|
79 |
+
for fold in folders2:
|
80 |
+
shutil.rmtree(fold, ignore_errors=True)
|
81 |
+
|
82 |
+
|
83 |
+
|
84 |
+
|
app.py
ADDED
@@ -0,0 +1,145 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from api import *
|
3 |
+
from processing import *
|
4 |
+
import pandas as pd
|
5 |
+
from indices import indices
|
6 |
+
import xgboost as xgb
|
7 |
+
#from lightgbm import LGBMRegressor
|
8 |
+
from sklearn.preprocessing import StandardScaler
|
9 |
+
from sklearn.decomposition import PCA
|
10 |
+
import pickle as pk
|
11 |
+
import json
|
12 |
+
import boto3
|
13 |
+
from shapely.geometry import MultiPolygon,shape
|
14 |
+
from shapely.geometry import Point
|
15 |
+
from shapely.geometry.polygon import Polygon
|
16 |
+
from glob import glob
|
17 |
+
import wget
|
18 |
+
|
19 |
+
|
20 |
+
def predict(location_name,lat, lon):
|
21 |
+
cord = [lon,lat]
|
22 |
+
lon = round(lon,4)
|
23 |
+
lat = round(lat,4)
|
24 |
+
x1 = [lon,lat]
|
25 |
+
x2 = [lat,lon]
|
26 |
+
with open("data/CIV_0.json","r") as file:
|
27 |
+
data = json.load(file)
|
28 |
+
# extract ivory coast polygone
|
29 |
+
features = [data['features'][0]['geometry']['coordinates'][0]+data['features'][0]['geometry']['coordinates'][1]+data['features'][0]['geometry']['coordinates'][2]]
|
30 |
+
data['features'][0]['geometry']['coordinates'] = features
|
31 |
+
ci_polygone = data['features'][0]['geometry']['coordinates'][0][0]
|
32 |
+
point1 = Point(x1)
|
33 |
+
point2 = Point(x2)
|
34 |
+
polygon = Polygon(ci_polygone)
|
35 |
+
result = polygon.contains(point1)
|
36 |
+
|
37 |
+
if not result:
|
38 |
+
return "Choisissez une zone de la CI","","","",""
|
39 |
+
|
40 |
+
else:
|
41 |
+
df = pd.read_csv("data/frame.csv")
|
42 |
+
name = find_good_tile(df,point2)
|
43 |
+
if name ==404:
|
44 |
+
reponse = "Sentinel-2 ne dispose pas de données ce sur ce lieu à ce jour"
|
45 |
+
return reponse,"","","",""
|
46 |
+
else:
|
47 |
+
path = "https://data354-public-assets.s3.eu-west-3.amazonaws.com/cisentineldata/"
|
48 |
+
url = path+name
|
49 |
+
#wget.download(url)
|
50 |
+
unzip()
|
51 |
+
name,cld_prob,days_ago = select_best_cloud_coverage_tile()
|
52 |
+
bandes_path_10,bandes_path_20,bandes_path_60,tile_path,path_cld_20,path_cld_60 =paths(name)
|
53 |
+
# create image dataset
|
54 |
+
images_10 = extract_sub_image(bandes_path_10,tile_path,cord)
|
55 |
+
|
56 |
+
# bandes with 20m resolution
|
57 |
+
#path_cld_20
|
58 |
+
images_20 = extract_sub_image(bandes_path_20,tile_path,cord,20,1)
|
59 |
+
|
60 |
+
# bandes with 60m resolution
|
61 |
+
#path_cld_60
|
62 |
+
images_60 = extract_sub_image(bandes_path_60,tile_path,cord,60)
|
63 |
+
#
|
64 |
+
feature = images_10.tolist()+images_20.tolist()+images_60.tolist()
|
65 |
+
bands = ['B02', 'B03', 'B04', 'B05', 'B06', 'B07', 'B08', 'B8A', 'B11', 'B12','B01','B09']
|
66 |
+
print("feature : ",feature)
|
67 |
+
print("BANDES : ",bands)
|
68 |
+
|
69 |
+
X = pd.DataFrame([feature],columns = bands)
|
70 |
+
print("==================== X SHAPE", X.shape)
|
71 |
+
## Coordinate
|
72 |
+
cord_df = pd.DataFrame({"Latitude":[lat],
|
73 |
+
"Longitude":[lon]})
|
74 |
+
print("==================== cord_df SHAPE", cord_df.shape)
|
75 |
+
## PCA dimension reduction
|
76 |
+
# later reload the pickle file
|
77 |
+
sdc_reload = pk.load(open("data/sdc.pkl",'rb'))
|
78 |
+
pca_reload = pk.load(open("data/pca.pkl",'rb'))
|
79 |
+
# standardization
|
80 |
+
X_pca = sdc_reload.transform(X)
|
81 |
+
# make pca
|
82 |
+
principalComponents = pca_reload .transform(X_pca)
|
83 |
+
principalDf = pd.DataFrame(data =principalComponents[:,:4],
|
84 |
+
columns = ["PC1","PC2","PC3","PC4"])
|
85 |
+
print("==================== principalDf SHAPE", principalDf.shape)
|
86 |
+
|
87 |
+
# vegetation index calculation
|
88 |
+
X = indices(X)
|
89 |
+
# Drop all 12 bands of S2
|
90 |
+
tab = list(range(12))
|
91 |
+
X_index = X.drop(X.iloc[:,tab],axis=1)
|
92 |
+
|
93 |
+
print("=============SHAPE1",X_index.shape)
|
94 |
+
|
95 |
+
# Create predictive features
|
96 |
+
X_final =pd.concat([cord_df,principalDf,X_index],axis=1)
|
97 |
+
print("=============SHAPE2",X_final.shape)
|
98 |
+
|
99 |
+
# load the model from disk
|
100 |
+
filename = "data/finalized_model3.sav"
|
101 |
+
loaded_model = pk.load(open(filename, 'rb'))
|
102 |
+
# make prediction
|
103 |
+
biomass = loaded_model.predict(X_final)[0]
|
104 |
+
if biomass<0:
|
105 |
+
biomass =0.0
|
106 |
+
|
107 |
+
carbon = 0.55*biomass
|
108 |
+
|
109 |
+
# NDVI
|
110 |
+
ndvi_index = ndvi(cord,name)
|
111 |
+
|
112 |
+
# deleted download files
|
113 |
+
#delete_tiles()
|
114 |
+
|
115 |
+
return str(cld_prob)+ " % cloud coverage", str(days_ago)+" days ago",str(biomass)+" t/ha", str(carbon)+" tC/ha","NDVI: "+ str(ndvi_index)
|
116 |
+
|
117 |
+
# Create title, description and article strings
|
118 |
+
title = "🌴BEEPAS : Biomass estimation to Evaluate the Environmental Performance of Agroforestry Systems🌴"
|
119 |
+
description = "This application estimates the biomass of certain areas using AI and satellite images (S2)."
|
120 |
+
article = "Created by data354."
|
121 |
+
|
122 |
+
# Create examples list from "examples/" directory
|
123 |
+
#example_list = [["examples/" + example] for example in os.listdir("examples")]
|
124 |
+
example_list = [["Foret du banco :",5.379913, -4.050445],["Pharmacie Y4 :",5.363292, -3.9481601],["Treichville Bernabé :",5.293168, -3.999796],["Adjamé :",5.346938, -4.027849],["ile boulay :",5.280498,-4.089883]]
|
125 |
+
|
126 |
+
|
127 |
+
outputs = [
|
128 |
+
gr.Textbox(label="Cloud coverage"),
|
129 |
+
gr.Textbox(label="Number of days since sensing"),
|
130 |
+
gr.Textbox(label="Above ground biomass density(AGBD) t/ha"),
|
131 |
+
gr.Textbox(label="Carbon stock density tC/ha "),
|
132 |
+
gr.Textbox(label="Mean NDVI"),]
|
133 |
+
|
134 |
+
|
135 |
+
demo = gr.Interface(
|
136 |
+
fn=predict,
|
137 |
+
inputs=["text","number", "number"],
|
138 |
+
outputs=outputs, #[ "text", "text","text","text","text"],
|
139 |
+
examples=example_list,
|
140 |
+
title=title,
|
141 |
+
description=description,
|
142 |
+
article=article,
|
143 |
+
)
|
144 |
+
|
145 |
+
demo.launch(share=True)
|