Basic Usage

Install¶

Installation is easy just grab it form pip

In [ ]:

!pip install virtughan==0.7.0

!pip install virtughan==0.7.0

Setup¶

Define the parameters for your computation

Use this for visual image timeseries

formula="band1"
band1="visual"
band2=None
operation=None
timeseries=True

In [9]:

# Define the bounding box
bbox =  [-17.942562103271488,28.569682805210718,-17.839050292968754,28.64689695054486] # cumbre vieja volcano spain area

# Define the parameters
start_date = "2021-10-01"
end_date = "2021-12-30"
cloud_cover = 10
formula = "(band1-band2)/(band1+band2)"  # NBR
band1 = "nir08"
band2 = "swir22"
operation = "max"
output_dir = "virtughan_output"
timeseries = True
workers = 1 # no of parallel workers
smart_filter=False

# Define the bounding box bbox = [-17.942562103271488,28.569682805210718,-17.839050292968754,28.64689695054486] # cumbre vieja volcano spain area # Define the parameters start_date = "2021-10-01" end_date = "2021-12-30" cloud_cover = 10 formula = "(band1-band2)/(band1+band2)" # NBR band1 = "nir08" band2 = "swir22" operation = "max" output_dir = "virtughan_output" timeseries = True workers = 1 # no of parallel workers smart_filter=False

Cleanup¶

Lets clear if there are previous output in the dir

In [10]:

import shutil , os
if os.path.exists(output_dir):
    shutil.rmtree(output_dir)

import shutil , os if os.path.exists(output_dir): shutil.rmtree(output_dir)

Example compute on time dimension¶

You can run computation now , You can visualize the results or download them from the files directly after computation is done

In [11]:

from virtughan.engine import VirtughanProcessor

processor = VirtughanProcessor(
    bbox,
    start_date,
    end_date,
    cloud_cover,
    formula,
    band1,
    band2,
    operation,
    timeseries,
    output_dir,
    workers=workers,
    smart_filter=smart_filter,
    cmap="RdYlGn"
)
processor.compute()

from virtughan.engine import VirtughanProcessor processor = VirtughanProcessor( bbox, start_date, end_date, cloud_cover, formula, band1, band2, operation, timeseries, output_dir, workers=workers, smart_filter=smart_filter, cmap="RdYlGn" ) processor.compute()

Engine starting...
Searching STAC .....
Total scenes found: 11
Scenes covering input area: 11
Scenes after removing overlaps: 6
Computing Band Calculation: 100%|██████████| 6/6 [00:07<00:00,  1.27s/it]
Aggregating results...
Saving aggregated result with colormap...
Creating GIF and zipping TIFF files...
Saved timeseries GIF to virtughan_output/output.gif
Saved intermediate images ZIP to virtughan_output/tiff_files.zip

In [12]:

from IPython.display import Image
Image(open(os.path.join(output_dir,'output.gif'),'rb').read())

from IPython.display import Image Image(open(os.path.join(output_dir,'output.gif'),'rb').read())

Out[12]:

Example on the fly tile computation¶

Here computation will be done alike google earth engine , you zoom out do computation , zoom in do computation based on your formula and resolution will differ accordingly as resampling will happen

In [15]:

import mercantile
from PIL import Image
from io import BytesIO
from virtughan.tile import TileProcessor

# Define the parameters
lat, lon = 28.28139, 83.91866
zoom_level = 12
x, y, z = mercantile.tile(lon, lat, zoom_level)

tile_processor = TileProcessor()

image_bytes, feature = await tile_processor.cached_generate_tile(
    x=x,
    y=y,
    z=z,
    start_date="2020-01-01",
    end_date="2025-01-01",
    cloud_cover=30,
    band1="red",
    band2="nir",
    formula="(band2-band1)/(band2+band1)",
    colormap_str="RdYlGn",
)

image = Image.open(BytesIO(image_bytes))

print(f"Tile: {x}_{y}_{z}")
print(f"Date: {feature['properties']['datetime']}")
print(f"Cloud Cover: {feature['properties']['eo:cloud_cover']}%")

image.save(f'tile_{x}_{y}_{z}.png')

import mercantile from PIL import Image from io import BytesIO from virtughan.tile import TileProcessor # Define the parameters lat, lon = 28.28139, 83.91866 zoom_level = 12 x, y, z = mercantile.tile(lon, lat, zoom_level) tile_processor = TileProcessor() image_bytes, feature = await tile_processor.cached_generate_tile( x=x, y=y, z=z, start_date="2020-01-01", end_date="2025-01-01", cloud_cover=30, band1="red", band2="nir", formula="(band2-band1)/(band2+band1)", colormap_str="RdYlGn", ) image = Image.open(BytesIO(image_bytes)) print(f"Tile: {x}_{y}_{z}") print(f"Date: {feature['properties']['datetime']}") print(f"Cloud Cover: {feature['properties']['eo:cloud_cover']}%") image.save(f'tile_{x}_{y}_{z}.png')

Tile: 3002_1712_12
Date: 2024-12-30T05:10:56.883000Z
Cloud Cover: 1.140427%

In [16]:

from IPython.display import Image as display_image
display_image(f'tile_{x}_{y}_{z}.png')

from IPython.display import Image as display_image display_image(f'tile_{x}_{y}_{z}.png')

Out[16]:

Extract Sentinel-2 Images for your AOI¶

This example showcase how you can select the band you want and extract sentinel-2 images just for your area of interest

In [17]:

import os

bbox = [83.84765625, 28.22697003891833, 83.935546875, 28.304380682962773]
start_date = "2024-12-15"
end_date = "2024-12-31"
cloud_cover = 30
bands_list = ["blue","green","red", "nir","swir16","swir22"] # order of the bands will result order in the resulting output
output_dir = "./sentinel_images"
workers = 1

os.makedirs(output_dir, exist_ok=True)

import os bbox = [83.84765625, 28.22697003891833, 83.935546875, 28.304380682962773] start_date = "2024-12-15" end_date = "2024-12-31" cloud_cover = 30 bands_list = ["blue","green","red", "nir","swir16","swir22"] # order of the bands will result order in the resulting output output_dir = "./sentinel_images" workers = 1 os.makedirs(output_dir, exist_ok=True)

In [18]:

from virtughan.extract import ExtractProcessor, VALID_BANDS

print("Available bands: ", VALID_BANDS)

extractor = ExtractProcessor(
    bbox,
    start_date,
    end_date,
    cloud_cover,
    bands_list,
    output_dir,
    workers=workers,
)
extractor.extract()

from virtughan.extract import ExtractProcessor, VALID_BANDS print("Available bands: ", VALID_BANDS) extractor = ExtractProcessor( bbox, start_date, end_date, cloud_cover, bands_list, output_dir, workers=workers, ) extractor.extract()

Available bands:  {'red': 'Red - 10m', 'green': 'Green - 10m', 'blue': 'Blue - 10m', 'nir': 'NIR 1 - 10m', 'swir22': 'SWIR 2.2μm - 20m', 'rededge2': 'Red Edge 2 - 20m', 'rededge3': 'Red Edge 3 - 20m', 'rededge1': 'Red Edge 1 - 20m', 'swir16': 'SWIR 1.6μm - 20m', 'wvp': 'Water Vapour (WVP)', 'nir08': 'NIR 2 - 20m', 'aot': 'Aerosol optical thickness (AOT)', 'coastal': 'Coastal - 60m', 'nir09': 'NIR 3 - 60m'}
Extracting bands...
Total scenes found: 4
Scenes covering input area: 4
Scenes after removing overlaps: 4
Filter from : 2024-12-15 to : 2024-12-30
Selecting 1 image per 4 days
Scenes after applying smart filter: 4
Extracting Bands:   0%|          | 0/4 [00:00<?, ?it/s]Stacking Bands...
Extracting Bands:  25%|██▌       | 1/4 [00:09<00:28,  9.67s/it]Stacking Bands...
Extracting Bands:  50%|█████     | 2/4 [00:19<00:19,  9.55s/it]Stacking Bands...
Extracting Bands:  75%|███████▌  | 3/4 [00:29<00:09,  9.78s/it]Stacking Bands...
Extracting Bands: 100%|██████████| 4/4 [00:37<00:00,  9.28s/it]

In [18]: