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