Segment Anything Model for Geospatial Data¶

This notebook shows how to use segment satellite imagery using the Segment Anything Model (SAM) with a few lines of code.

Make sure you use GPU runtime for this notebook. For Google Colab, go to Runtime -> Change runtime type and select GPU as the hardware accelerator.

Install dependencies¶

Uncomment and run the following cell to install the required dependencies.

In [ ]:

# %pip install leafmap

# %pip install leafmap

In [ ]:

# %pip install segment-geospatial localtileserver

# %pip install segment-geospatial localtileserver

Import libraries¶

In [ ]:

import os
import leafmap
from samgeo import SamGeo, tms_to_geotiff, get_basemaps

import os import leafmap from samgeo import SamGeo, tms_to_geotiff, get_basemaps

Create an interactive map¶

In [ ]:

m = leafmap.Map(center=[29.676840, -95.369222], zoom=19)
m.add_basemap("SATELLITE")
m

m = leafmap.Map(center=[29.676840, -95.369222], zoom=19) m.add_basemap("SATELLITE") m

Pan and zoom the map to select the area of interest. Use the draw tools to draw a polygon or rectangle on the map

In [ ]:

if m.user_roi_bounds() is not None:
    bbox = m.user_roi_bounds()
else:
    bbox = [-95.3704, 29.6762, -95.368, 29.6775]

if m.user_roi_bounds() is not None: bbox = m.user_roi_bounds() else: bbox = [-95.3704, 29.6762, -95.368, 29.6775]

Download map tiles¶

Download maps tiles and mosaic them into a single GeoTIFF file

In [ ]:

image = "satellite.tif"

image = "satellite.tif"

Besides the satellite basemap, you can use any of the following basemaps returned by the get_basemaps() function:

In [ ]:

# get_basemaps().keys()

# get_basemaps().keys()

Specify the basemap as the source.

In [ ]:

tms_to_geotiff(output=image, bbox=bbox, zoom=20, source="Satellite", overwrite=True)

tms_to_geotiff(output=image, bbox=bbox, zoom=20, source="Satellite", overwrite=True)

You can also use your own image. Uncomment and run the following cell to use your own image.

In [ ]:

# image = '/path/to/your/own/image.tif'

# image = '/path/to/your/own/image.tif'

Display the downloaded image on the map.

In [ ]:

m.layers[-1].visible = False  # turn off the basemap
m.add_raster(image, layer_name="Image")
m

m.layers[-1].visible = False # turn off the basemap m.add_raster(image, layer_name="Image") m

Initialize SAM class¶

In [ ]:

sam = SamGeo(
    model_type="vit_h",
    checkpoint="sam_vit_h_4b8939.pth",
    sam_kwargs=None,
)

sam = SamGeo( model_type="vit_h", checkpoint="sam_vit_h_4b8939.pth", sam_kwargs=None, )

Segment the image¶

Set batch=True to segment the image in batches. This is useful for large images that cannot fit in memory.

In [ ]:

mask = "segment.tif"
sam.generate(
    image, mask, batch=True, foreground=True, erosion_kernel=(3, 3), mask_multiplier=255
)

mask = "segment.tif" sam.generate( image, mask, batch=True, foreground=True, erosion_kernel=(3, 3), mask_multiplier=255 )

Polygonize the raster data¶

Save the segmentation results as a GeoPackage file.

In [ ]:

vector = "segment.gpkg"
sam.tiff_to_gpkg(mask, vector, simplify_tolerance=None)

vector = "segment.gpkg" sam.tiff_to_gpkg(mask, vector, simplify_tolerance=None)

You can also save the segmentation results as any vector data format supported by GeoPandas.

In [ ]:

shapefile = "segment.shp"
sam.tiff_to_vector(mask, shapefile)

shapefile = "segment.shp" sam.tiff_to_vector(mask, shapefile)

Visualize the results¶

In [ ]:

style = {
    "color": "#3388ff",
    "weight": 2,
    "fillColor": "#7c4185",
    "fillOpacity": 0.5,
}
m.add_vector(vector, layer_name="Vector", style=style)
m

style = { "color": "#3388ff", "weight": 2, "fillColor": "#7c4185", "fillOpacity": 0.5, } m.add_vector(vector, layer_name="Vector", style=style) m