Detailed Analysis Capabilities

Heimdall provides deep, agronomic insights for every cell of your field. You choose how fine the analysis grid should be—from highly detailed small cells to larger sizes for vast fields. Discover how our on-device machine learning turns raw drone imagery into precise, actionable data.

Crop Growth & Vitality

Heimdall continuously tracks crop development by calculating Growing Degree Days (GDD) since sowing to estimate the precise BBCH growth stage. It measures 7-day height changes to identify stress, waterlogging, or stunted growth. Additionally, it calculates the Excess Green (ExG) index to monitor photosynthetic activity and overall canopy health.

Weed Detection & Species ID

Using localized machine learning, Heimdall identifies individual weed species and assigns them international EPPO codes. It calculates the Exponential Moving Average (EMA) of weed density per cell to filter out single-flight noise. Weeds are automatically categorized by herbicide group (grass vs. broadleaf), allowing you to formulate targeted tank mixes and combat resistance.

Fungal Disease Severity

The system detects visual symptoms of fungal infections (such as rust, septoria, or fusarium) directly on the crop canopy. It calculates the disease severity as the precise percentage of the affected canopy area, enabling variable-rate fungicide applications exactly where economic thresholds are crossed.

Insect Pest Monitoring

Heimdall identifies and counts pest insects down to the square meter. By comparing the smoothed EMA pest density against your pre-defined economic thresholds, the system ensures insecticides are only prescribed when the cost of crop damage outweighs the treatment cost.

Crop Lodging & Harvest Planning

Through specialized 15m oblique flights, Heimdall utilizes texture analysis to detect wind-blown or flattened crop areas (lodging). The Harvest Planner highlights these lodging risks—which can increase grain contamination and drying costs—allowing you to optimize your combine harvester's routing.

Supported Crop Profiles

Heimdall has dedicated agronomic profiles (growth models, BBCH phenology scaling, biomass proxies, and specific analysis routing) for the following primary row crops and broadacre crops:

  • Maize / Corn
  • Cereals (Wheat, Barley, Rye)
  • Potato
  • Oilseed Rape / Canola
  • Soybean
  • Sugar Beet
  • Sunflower

(If a crop isn't in this list, Heimdall falls back to a Generic Crop Profile which uses a generic BBCH scale and tracks a baseline "Canopy Vigor" index).

Recognized Species

The ML detection pipeline identifies species based on the European and Mediterranean Plant Protection Organization (EPPO) standards. The detections are categorized as follows:

Grass Weeds (Graminicide Targets)

  • Black-grass (ALOMY)
  • Wild oat (AVEFA), Animated oat (AVEST)
  • Perennial ryegrass (LOLPE), Italian ryegrass (LOLMU), Rigid ryegrass (LOLRI)
  • Annual meadow-grass (POAAN)
  • Barnyard-grass (ECHCG)
  • Green foxtail (SETVI)
  • Hairy crabgrass (DIGSA)
  • Loose silky bent (APESV)
  • Generic Grass Family (1POAF)

Broadleaf Weeds (Dicot Targets)

  • Fat-hen / Lambsquarters (CHEAL)
  • Creeping thistle (CIRAR)
  • Chickweed (STEME)
  • Scented mayweed (MATCH)
  • Common poppy (PAPRH)
  • Field bindweed (CONAR)
  • Charlock (SINAR), Wild radish (RRARA)
  • Field pennycress (THLAR)
  • Amaranth family: Palmer amaranth (AMAPA), Redroot pigweed (AMARE), Waterhemp (AMATU)
  • Ragweed (AMBAR)
  • Velvetleaf (ABUTH)
  • Kochia (BASSO)
  • Dandelion (TAROF), Sow-thistle (SONOL)
  • ...and over a dozen other regional speedwells, fleabanes, and soldiers.

Diseases & Fungi

  • Yellow rust (PUCCST)
  • Late blight (PHYPIN) - Potato
  • Septoria tritici blotch (SEPTTR) - Cereals
  • Fusarium head blight (GIBBZE)
  • Grey mould (BOTRCI)

Pests & Insects

  • Peach-potato aphid (MYZUPE)
  • Western corn rootworm (DIABVI)
  • Colorado potato beetle (LPTNDE)
  • Wireworm (AGRIXX)

Manure Spreading Analysis

By comparing drone imagery taken before and after application, Heimdall measures color shifts (ΔE in CIELAB space) to assess spreading uniformity. It accurately detects missed cells, overlapping applications, regulatory buffer zone violations near watercourses, and can even estimate the manure type (poultry, cattle, digestate, pig) based on its unique color signature.