Objectives of the service
Broccoli growers currently cannot predict harvest timing more than two to three days ahead. This forces reactive planning of harvest crews, machinery, and logistics, and leaves growers with little leverage in sales negotiations. The core problem is that crop development varies between and within fields in ways that cannot be tracked efficiently through manual field inspections alone.
QualityTime addresses this by combining daily satellite imagery from Planet SuperDove (3 metre resolution) with a temperature-driven crop growth model to detect how each field is developing relative to its expected harvest date, starting from early in the season. The result is a harvest window prediction delivered 30 or more days in advance, alongside a spatial map showing which zones within each field will be ready first. These outputs are delivered through a mobile-optimised platform, keeping information accessible to farmers and their crews in the field.
The KickStart activity set out to prove that this approach works in practice: engaging the broccoli farming community in North-Holland to define requirements and validate the concept, building and testing the prediction pipeline against real field data, and testing the first prototypes of the platform with one pilot grower, while assessing whether the service can be delivered commercially at a price growers are willing to pay.
Users and their needs
Growers of broccoli are always observing their crops to evaluate if the harvest can be executed as planned. Harvest is labour intensive, so timely planning is key. Also, broccoli is a fresh product, and is sold to e.g., retailers. Timely harvest information allows the value chain to absorb yield peaks. QualityTime delivers timely harvest predictions, both timing and volumes, to help growers in a more sustainable and profitable business.
QualityTime starts in the Netherlands but aims to serve broccoli growers around the world.
Service/ system concept
Space Added Value
The QualityTime service uses Planet SuperDove satellite imagery (3 m resolution, daily revisit) as its primary space asset, deriving WDVI as a continuous biomass proxy throughout the growing season, complemented by drone imagery for targeted field-level calibration. Current practice relies on manual field inspections and calendar-based predictions from seed breeders, which provide at most 2–3 days' advance notice and require labour-intensive scouting of every field individually. Planet SuperDove changes this in three fundamental ways: every registered field is monitored daily without any grower effort, making whole-farm portfolio monitoring economically viable at scale; the WDVI time series combined with a GDD crop growth model enables harvest window predictions 30+ days in advance, far beyond what any field-based method can offer; and the 3 m pixel size resolves crop variability within a field at the phenological inflection point, producing harvest zone maps that no competitor currently delivers at commercial scale. The combination of daily satellite coverage with targeted drone calibration, where the satellite maps guide which fields to fly and when, is the key differentiator: neither data source alone achieves the same predictive accuracy or operational efficiency.
Current Status
Status July 2026: broccoli are partly harvested. Fielddata is collected into our portal. Drone based validation has been done.
The technical feasibility of satellite remote sensing as an early predictor of harvest variability in broccoli has been validated. More than 30 growers and farm employees were interviewed and consulted to define system requirements. A working proof of concept has been built around one pilot grower's fields, which are actively tracked throughout the 2026 growing season. To calibrate the prediction model, over 20 hectares of high-resolution drone imagery was collected using a newly developed rapid-acquisition methodology, significantly faster than conventional approaches, and processed with a purpose-built AI head detection model. Analysis of this imagery is currently underway, comparing drone-derived ground truth against the system's early satellite-based harvest predictions to further refine and calibrate the model. Next steps include validation across additional pilot fields and initiating commercial onboarding discussions with three growers who have expressed interest in the service.
The kickstart activity is successfully concluded, and indicated that the service is technical feasible, economically viable, and realistically operational within next years.
Prime Contractor(s)
