Skip to article frontmatterSkip to article content

Segmentation Model

1Under development

2Part 1:

  • viewing the satellite imagery file of pabbi on interactive map
  • viewing the pabbi mauzas boundary.
  • viewing the overture map building information as a layer on the interactive map.
import leafmap.foliumap as leafmap
import geopandas as gpd
hayat_titiler = 'http://imhayatkhan-hayat-titiler.hf.space'
file_url = 'https://pabbiimagery.s3.eu-north-1.amazonaws.com/pabbi_satellte_image_COG.tif'
geojson_url = "https://raw.githubusercontent.com/code4geoai/gee/main/pabb_crop_V3.geojson"
fb_s3 = 's3://pabbiimagery/overturemaps_filteredbuildings.parquet'
fb_parquet_gdf = gpd.read_parquet(fb_s3, storage_options={"anon": True}) 


m = leafmap.Map(center=[34.0174, 71.7749], zoom=14)
m.add_basemap("Esri.WorldImagery")

# Add your COG layer using public Titiler endpoint
m.add_cog_layer(
    url=file_url,
    titiler_endpoint=hayat_titiler,  # 
    name="Pabbi COG (tiled)",
    bands=[1,2,3],       # Titiler will still rescale for you
   # rescale=["10,190","0,200","0,200"],  # rescale each band
    #opacity=0.7,
)

m.add_geojson(
    geojson_url,
    layer_name="Boundary",
    style={"color": "green", "weight": 0.5, "fillOpacity": 0.1,},
    visible=False,
)

m.add_geojson(
    fb_parquet_gdf,
    layer_name="Buildings",
    style={"color": "blue", "weight": 0.5, "fillOpacity": 0.1,},
    )
m
Loading...

3Part 2:

  • Creating image chips
  • Training the model on image chips
  • Applying the model.
  • ........Coming soon...