Skip to main content

FHI Overview

Field Health Imagery (FHI)​

FHI is an enrollment-based service.
Submit one enrollment to define:

  • field boundaries with customer-defined field_id values
  • monitoring window (start_date to end_date)
  • requested imagery layers

After submission, processing is asynchronous. FHI backfills historical dates in range and continues daily checks for new scenes during the monitoring window.

Interactive API specification​

For the live endpoint contract, request/response schemas, and try-it workflow, use the FHI Swagger UI:

Runtime behavior and expectations​

A completed enrollment means enrollment artifacts were ingested and scheduled — it does not guarantee imagery for every field-date. Imagery is produced when source imagery is available and passes quality checks; missing dates are expected in cases such as acquisition gaps or low-quality scenes.

See FHI Service Behavior & Expectations for expected imagery timelines, common scenarios and what they mean, what FHI guarantees vs. what's external, and field boundary update behavior.

Image quality and daily selection​

For each field-date, FHI evaluates available scenes and selects one representative image for processing. Selection is deterministic and quality-driven, keeping outputs consistent across analytical and visualization layers.

See FHI Service Behavior & Expectations — Image quality and daily selection for the actual quality thresholds and how filtering vs. selection works.

Output types​

FHI generates multiple output types, including analytics, visualization derivatives, and reference imagery.

Base analytics products​

  • ndvi.tiff, ndre.tiff, gndvi.tiff, evi.tiff, ccci.tiff, osavi.tiff, savi.tiff: vegetation index rasters for quantitative analysis
  • sr_rgb.tiff: surface reflectance RGB for analytical workflows
  • sr-multispectral.tiff: surface reflectance multispectral bands, for custom index computation or downstream analysis

Visualization derivatives​

For each supported index, FHI can generate:

  • *_vegetation_map.tiff: color-mapped absolute interpretation of vegetation health
  • *_scouting_map.tiff: relative stress/anomaly interpretation for field scouting
tip

Scouting maps are relative interpretations. Avoid direct comparison across different dates.

Reference imagery​

  • true_color.tiff: natural-color RGB context imagery

Output storage and access​

Generated outputs are written to a deterministic storage path:

{customer}/fhi/planet/{field_id}/{date}/{image_type}.tiff

Example: bayer/fhi/planet/FIELD_001/2025-02-15/ndvi.tiff

See Accessing Your Imagery for a step-by-step walkthrough of requesting a token and browsing/downloading files with a storage tool.

Enrollment scale and limits​

Documented operational limits for this release cycle:

  • maximum GeoJSON file size: 50 MB
  • maximum fields per enrollment: 250

For high-volume onboarding, validate a representative sample enrollment before production rollout.