Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.wherobots.com/llms.txt

Use this file to discover all available pages before exploring further.

Private Preview
The following content is a read-only preview of an executable Jupyter notebook.To run this notebook interactively:
  1. Go to the Wherobots Model Hub.
  2. Select the specific notebook you wish to run.
  3. Click Run Model in Notebook.
chm banner This notebook will guide you through estimating canopy height from aerial imagery, using Wherobots RasterFlow and the Meta and World Resources Institute’s Canopy Height prediction model (Meta CHM v1)1. This is an open source regression model that can predict the height of the tree canopy from high resolution imagery. The result of the model is a raster where each pixel is the estimated tree canopy height in meters. It was trained on high-resoluton imagery (Maxar Vivid2 mosaic with 0.5m resolution), existing labeled canopy height maps and airborne laser scans (LIDAR).

​
Running inference on high resolution imagery

The Meta CHM v1 model was trained on 50cm resolution data, so we will select imagery with a similar resolution for our testing. The National Agriculture Imagery Program (NAIP) provides aerial imagery for the United States, capturing high-resolution images during the agricultural growing seasons. The NAIP dataset contains a mixture of resolutions (30cm, 60cm and 1m). See this map for more details. We will use 60cm imagery to test this model.

​
Selecting an Area of Interest (AOI)

We will choose an Area of Interest (AOI) for our analysis. The area around Nashua, NH has a combination of urban settings, parks and forests so we will try out the model there. From the map, we see that 60cm imagery for New Hampshire was last captured in 2021, so we will set our time range accordingly. 
import wkls
import geopandas as gpd
import os

# Generate a geometry for Nashua, NH using WKLS (https://github.com/wherobots/wkls)
gdf = gpd.read_file(wkls.us.nh.nashua.geojson())

# Save the geometry to a parquet file in the user's S3 path
aoi_path = os.getenv("USER_S3_PATH") + "nashua.parquet"
gdf.to_parquet(aoi_path)

# we'll make variables for the bounds of our aoi for use after running inference to visualize results
min_lon, min_lat, max_lon, max_lat = gdf.total_bounds

​
Initializing the RasterFlow client

from datetime import datetime

from rasterflow_remote import RasterflowClient

from rasterflow_remote.data_models import (
    ModelRecipes
)

rf_client = RasterflowClient()

​
Running a model

RasterFlow has pre-defined workflows to simplify orchestration of the processing steps for model inference. These steps include:
  • Ingesting imagery for the specified Area of Interest (AOI)
  • Generating a seamless image from multiple image tiles (a mosaic)
  • Running inference with the selected model
The output is a Zarr store of the model outputs. Note: This step will take approximately 20 minutes to complete. 
model_output_index = rf_client.build_and_predict_mosaic_recipe(
    # Path to our AOI in GeoParquet or GeoJSON format
    aoi = aoi_path,

    # Date range for imagery to be used by the model
    start = datetime(2021, 1, 1),
    end = datetime(2022, 1, 1),

    # Coordinate Reference System EPSG code for the output mosaic   
    crs_epsg = 3857,

    # The model recipe to be used for inference (CHM in this case)
    model_recipe = ModelRecipes.META_CHM_V1
)

model_output_index.mosaics

# This example AOI has one geometry, so we select the first (only) mosaic location.
model_output_store = model_output_index.first_row_mosaic

model_output_store

​
(Optional) Build an optimized Zarr for visualization

RasterFlow writes its outputs as Zarr stores at native resolution. To explore them interactively on cloud.wherobots.com/map, you can build an optimized multiscale Zarr. build_zarr_multiscales adds downsampled overview levels (image pyramids) to the store so the map can stream coarse tiles when zoomed out and full-resolution pixels when zoomed in. This step is optional and can take a few minutes for large outputs, so the code below is commented out by default — uncomment it to run it. 
# optimized_store = rf_client.build_zarr_multiscales(source_store=model_output_store)
# optimized_store

​
Visualizing outputs

If RasterFlow is enabled for your organization, you can visualize the Zarr, GeoParquet, and other geospatial outputs using cloud.wherobots.com/map.

​
References

  1. Tolan, J., Yang, H.-I., Nosarzewski, B., Couairon, G., Vo, H. V., Brandt, J., Spore, J., Majumdar, S., Haziza, D., Vamaraju, J., et al. (2024). Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar. Remote Sensing of Environment, 300, 113888.