Get Started with RasterFlow

Public Preview

Get Started

Get started running RasterFlow in Wherobots.

Roof Inspection Guide

Inspect a property portfolio for roofs and solar panels in one SAM3 run.

Detect Objects with SAM3

Detect objects in imagery from natural language text prompts.

RasterFlow Datasets

Learn about built-in datasets and how to bring your own.

RasterFlow Models

Learn about built-in models and how to bring your own.

Build Multiscales

Add overview levels to a Zarr store so it can be visualized interactively.

RasterFlow Visualization

View mosaics, predictions, and vectorized results on a map.

Storage Integration

Read and write RasterFlow results directly in your own S3 bucket.

Run a Task

Run RasterFlow from any script or service through the Tasks API.

RasterFlow Billing

Learn how RasterFlow usage is metered in RasterFlow Spatial Units.

Run RasterFlow Tasks

Run tasks and read their results with the Python SDK or REST.

RasterFlow replaces the previous Raster Inference feature. For more information, see the changelog.

RasterFlow is the Wherobots image preparation and inference engine for large-scale raster processing and geospatial machine learning workflows. It builds mosaics from satellite and aerial imagery, runs computer vision models over them at scale, and vectorizes the predictions for analysis in WherobotsDB.

Why choose RasterFlow?

RasterFlow runs geospatial raster processing tasks as managed workflows. It abstracts away the complexity of raster pipelines and distributed computing, allowing you to focus on your analysis rather than data engineering and infrastructure management. Key capabilities include:

Planetary-scale processing

Process raster data at massive scale with optimized chunking, sharding, and parallel processing

Simple, high-level API

Abstract away the complexity with pre-configured datasets and models—or bring your own datasets and models.

Build mosaics

Combine multiple raster datasets into unified mosaics that are ready for further analysis

Run model inference

Apply machine learning models to massive raster datasets at scale

Vectorize results

Convert raster predictions into vector geometries for spatial analysis

Build Multiscales

Add overview levels to mosaics and model outputs so they can be viewed interactively on a map

Custom workflows

Build flexible pipelines with your own data, models, and processing parameters

Standard format support

Work with GeoTIFF, Zarr, and GeoParquet formats

Key concepts

The following concepts are fundamental to understanding how RasterFlow works:

Mosaics

Mosaics are spatially-aligned raster datasets stored in Zarr format. RasterFlow can build mosaics from aerial and satellite imagery to prepare them for model inference. This process can combine one or more datasets across an Area of Interest (AOI) and a temporal dimension to create a single seamless input mosaic for inference.RasterFlow has tasks for building mosaics from the built-in datasets (build_mosaics) or your own imagery (build_gti_mosaics).

Model inference

Run computer vision models on raster data at scale:

Semantic Segmentation: Classify each pixel (e.g., land cover mapping)
Regression: Predict continuous values (e.g., canopy height estimation in meters)
Patch-based Processing: Handle large mosaics by dividing into manageable patches

Vectorization

Convert raster predictions to vector geometries:

Threshold-based: Binarize continuous predictions
Polygonization: Create polygon features from classified pixels
Coordinate transformation: Reproject to desired CRS (e.g., WGS84)

Build Multiscales

RasterFlow writes mosaics and model outputs as Zarr stores at a single native resolution, which suits analysis but makes interactive viewing slow. Building multiscales reads an existing store and writes a new one that adds downsampled overview levels—an image pyramid—plus histogram statistics, so a map client can stream coarse tiles when you are zoomed out and full-resolution pixels when you zoom in. The source store is left unchanged.Build Multiscales with RasterFlow

Get started

There are two ways to run RasterFlow, and both run the same tasks: call the Tasks API from Python or any HTTP client, or open a ready-made notebook from the Models page.

Python SDK
Notebook

Every RasterFlow workflow is a task in the Tasks API. Start a task run, poll it until it finishes, then read its output URI from results. Inside a Wherobots notebook or job, the Wherobots Python SDK uses your Wherobots credentials. Anywhere else, set a Wherobots API key in WHEROBOTS_API_KEY. The SDK needs Python 3.11 or later.

# pip install wherobots-python-sdk
import time

import wherobots

wb = wherobots.Wherobots()  # outside Wherobots, reads WHEROBOTS_API_KEY

started = wb.tasks.run("predict_mosaic_geometries_recipe", {
    "aoi": "s3://wherobots-examples/rasterflow/aois/collegepark.parquet",
    "start": "2023-01-01T00:00:00Z",
    "end": "2023-12-31T00:00:00Z",
    "model_recipe": "sam3-text-geometry",
    "text_prompt": ["roofs"],
    "confidence_threshold": 0.3,
}, name="College Park roofs", region="aws-us-west-2")

run = wb.task_runs.get(started.task_run_id)
while run.is_active:
    time.sleep(30)
    run = wb.task_runs.get(run.task_run_id)

print(run.state, run.results, run.error_message)

curl -X POST "https://api.cloud.wherobots.com/tasks/predict_mosaic_geometries_recipe/runs" \
  -H "X-API-Key: $WHEROBOTS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "region": "aws-us-west-2",
    "name": "College Park roofs",
    "inputs": {
      "aoi": "s3://wherobots-examples/rasterflow/aois/collegepark.parquet",
      "start": "2023-01-01T00:00:00Z",
      "end": "2023-12-31T00:00:00Z",
      "model_recipe": "sam3-text-geometry",
      "text_prompt": ["roofs"],
      "confidence_threshold": 0.3
    }
  }'

# Repeat with the taskRunId from the response until state is no longer
# PENDING, SUBMITTED, RUNNING or CANCELLING.
curl "https://api.cloud.wherobots.com/task-runs/<taskRunId>" \
  -H "X-API-Key: $WHEROBOTS_API_KEY"

For regions, run states and each task’s output, see Run RasterFlow Tasks.

Next steps

Try out the pre-configured model solutions notebooks on the Models page in Wherobots Cloud
Run RasterFlow notebooks from your own editor with the Wherobots VS Code Extension
Run RasterFlow from any script or service through the Tasks API for automated, production-scale processing
Learn how RasterFlow usage is metered in RasterFlow Spatial Units
See Run RasterFlow Tasks for each task’s inputs, outputs, and run states

API reference

For detailed API documentation, see:

Run a task - POST /tasks/{taskName}/runs
Get a task run - state, results, and errors
Wherobots Python SDK - wb.tasks and wb.task_runs

Get Started with RasterFlow

Get Started

Roof Inspection Guide

Detect Objects with SAM3

RasterFlow Datasets

RasterFlow Models

Build Multiscales

RasterFlow Visualization

Storage Integration

Run a Task

RasterFlow Billing

Run RasterFlow Tasks

Why choose RasterFlow?

Key concepts

Get started

Text-Prompted Object Detection

Agricultural Field Mapping

Urban Infrastructure Detection

Canopy Height Estimation

Rural Road Detection

Next steps

API reference

Get Started

Roof Inspection Guide

Detect Objects with SAM3

RasterFlow Datasets

RasterFlow Models

Build Multiscales

RasterFlow Visualization

Storage Integration

Run a Task

RasterFlow Billing

Run RasterFlow Tasks

​Why choose RasterFlow?

​Key concepts

​Get started

Text-Prompted Object Detection

Agricultural Field Mapping

Urban Infrastructure Detection

Canopy Height Estimation

Rural Road Detection

​Next steps

​API reference

Why choose RasterFlow?

Key concepts

Get started

Next steps

API reference