Skip to main content
This example demonstrates query inference using a classification model with WherobotsAI Raster Inference to identify land cover in satellite imagery. We will use a machine learning model from torchgeo1 trained using imagery from the European Space Agency’s Sentinel-2 satellites. This model categorizes land cover into 10 categories from the EuroSat dataset 2:
  • Annual Crop
  • Forest
  • Herbaceous Vegetation
  • Highway
  • Industrial Buildings
  • Pasture
  • Permanent Crop
  • Residential Buildings
  • River
  • SeaLake

​
Before you start

This is a read-only preview of this notebook. To execute the cells in this Jupyter Notebook, do the following:
  1. Login to Wherobots Cloud.
  2. Start a GPU-Optimized runtime instance.
  3. Open a notebook.
      • We recommend using a Tiny GPU-Optimized runtime.
  4. Open the examples/Analyzing_Data/Raster_Classification.ipynb notebook path.
For more information on starting and using notebooks, see the following Wherobots Documentation:

​
Access a GPU-Optimized runtime

This notebook requires a GPU-Optimized runtime. For more information on GPU Optimized runtimes, see Runtime types. To access this runtime category, do the following:
  1. Sign up for a paid Wherobots Organization Edition (Professional or Enterprise).
  2. Submit a Compute Request for a GPU-Optimized runtime.

​
Step 1: Set up the Wherobots Context

import warnings
warnings.filterwarnings('ignore')

from wherobots.inference.data.io import read_raster_table
from sedona.spark import SedonaContext
from pyspark.sql.functions import expr

config = SedonaContext.builder().appName('classification-batch-inference')\
    .getOrCreate()

sedona = SedonaContext.create(config)

​
Step 2: Load satellite imagery

Next, we load the satellite imagery that we will be running inference over. These GeoTiff images are loaded as out-db rasters in WherobotsDB, where each row represents a different scene. 
tif_folder_path = 's3a://wherobots-examples/data/eurosat_small'
files_df = read_raster_table(tif_folder_path, sedona)
df_raster_input = files_df.withColumn(
        "outdb_raster", expr("RS_FromPath(path)")
    )
df_raster_input.cache().count()
df_raster_input.show(truncate=False)

​
Step 3: Run prediction with sedona.sql apis

To run predictions we will specify the model we wish to use. Some models are pre-loaded and made available in Wherobots Cloud. Predictions can be run using the Raster Inference SQL function RS_CLASSIFY or with the Python API. Here we generate 200 predictions using RS_CLASSIFY. 
%%time
df_raster_input.createOrReplaceTempView("df_raster_input")
model_id = 'landcover-eurosat-sentinel2'
predictions_df = sedona.sql(f"SELECT name, outdb_raster, RS_CLASSIFY('{model_id}', outdb_raster) AS preds FROM df_raster_input")
predictions_df.cache().count()
predictions_df.show(truncate=False)
predictions_df.createOrReplaceTempView("predictions_df")
You can specify your own model instead of using one of our hosted models via the model_id variable. To do so, replace the model_id variable with the s3 uri pointing to your Machine Learning Model Extension (MLM) metadata json. Then pass that as an argument to RS_CLASSIFY. For example: 
user_mlm_uri = 's3://wherobots-modelhub-prod/community/classification/landcover-eurosat-sentinel2/model-metadata.json'
predictions_df = sedona.sql(f"SELECT name, outdb_raster, RS_CLASSIFY('{user_mlm_uri}', outdb_raster) AS preds FROM df_raster_input")
Learn more about bringing your own model see Bring your own model in the Wherobots Documentation. From the prediction result, we can retrieve the most confidence classification label and it’s probability score. 
max_predictions_df = sedona.sql(f"SELECT name, outdb_raster, RS_MAX_CONFIDENCE(preds).max_confidence_label, RS_MAX_CONFIDENCE(preds).max_confidence_score FROM predictions_df")
max_predictions_df.show(20, truncate=False)

​
wherobots.inference Raster Inference Python API

If you prefer Python, wherobots.inference offers a module for registering the SQL inference functions as python functions. Below we run the same inference as before with RS_CLASSIFY. 
from wherobots.inference.engine.register import create_single_label_classification_udfs
rs_classify, rs_max_confidence = create_single_label_classification_udfs(batch_size = 10, sedona=sedona)
df_predictions = df_raster_input.withColumn("preds", rs_classify(model_id, 'outdb_raster'))
df_predictions.show(1)

from pyspark.sql.functions import col

df_max_predictions = df_predictions.withColumn("max_confidence_temp", rs_max_confidence(col("preds"))) \
                            .withColumn("max_confidence_label", col("max_confidence_temp.max_confidence_label")) \
                            .withColumn("max_confidence_score", col("max_confidence_temp.max_confidence_score")) \
                            .drop("max_confidence_temp", "preds")
df_max_predictions.cache().count()
df_max_predictions.show(2, truncate=False)

​
Visualize the model predictions and source imagery

df_rast = sedona.read.format("binaryFile").option("pathGlobFilter", "*.tif").option("recursiveFileLookup", "true").load(tif_folder_path).selectExpr("RS_FromGeoTiff(content) as raster")

htmlDF = df_max_predictions.selectExpr("RS_Band(outdb_raster, Array(4, 3, 2)) as image_raster", "name", "max_confidence_label")\
    .selectExpr("RS_NormalizeAll(image_raster, 1, 65535, True) as image_raster", "name", "max_confidence_label")\
    .selectExpr("RS_AsImage(image_raster, 500) as image_raster", "name", "max_confidence_label")

from sedona.raster_utils.SedonaUtils import SedonaUtils
from pyspark.sql.functions import rand
SedonaUtils.display_image(htmlDF.orderBy(rand()).limit(3))

​
References

  1. Stewart, A. J., Robinson, C., Corley, I. A., Ortiz, A., Lavista Ferres, J. M., & Banerjee, A. (2022). TorchGeo: Deep Learning With Geospatial Data. In Proceedings of the 30th International Conference on Advances in Geographic Information Systems (pp. 1-12). Association for Computing Machinery. https://doi.org/10.1145/3557915.3560953
  2. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. Patrick Helber, Benjamin Bischke, Andreas Dengel, Damian Borth. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019.