FRAVEBOT

Yield forecasting · for tomatoes, more high-wire crops coming

Fewer surprises at harvest: a forecast from every fruit.

Scout tracks every fruit and its colour from pass to pass. Yield Forecast turns that into kg/m² 7 and 14 days ahead, with a range – and shows how well it did last time.

High-wire tomato trusses ripening from green to red, every tomato and every cluster marked by Scout
Week 46 · forecast1.14 kg/m²
Yield Forecast chart, weeks 37 to 47: harvested kg/m² per week, the forecast 1.14 kg/m² for week 46 and 0.69 for week 47, with the greenhouse temperature drawn as an orange line above the bars
— °C greenhouse · warmer W46 → more
Forecast 7 days ahead89 %accurate
Forecast 14 days ahead87 %accurate
Fresh every morning4:30new forecast in the app and on WhatsApp

Accuracy compares the fruit we forecast as ripe, by a defined ripeness stage, with the fruit actually harvested. Not every ripe fruit is always picked, so harvest labour errors are part of the difference.

In the application

Every week, checked against what you picked

Dark bars: harvested. Blue bars: the forecast, with its likely range. Below, how far each week was off, so you know how much to trust it.

From week 41 the forecast ran low. Common at the end of the season, when growers pick faster than the crop’s pace, for example while clearing the greenhouse. The model cannot see those decisions coming.

Forecast chart in the application over 8 weeks: harvested kg/m² for weeks 37 to 44 with the forecast made before each week, the forecast for weeks 45 to 47 with a likely range, and below it the forecast accuracy, 4 of 6 weeks within plus or minus 15 percent

What goes in

The robot's data, your climate and your harvest, in one forecast.

From the robot

Plant data from Scout

  • Fruit coloursEvery fruit on a 0–70 colour scale, tracked from pass to pass.
  • Fruit defectsBlossom end rot, fruit defects and botrytis, taken out of the yield.
From the greenhouse

Greenhouse climate

  • Climate computerTemperature and light from the climate computer, plus the weather forecast.PrivaHoogendoornRidderand others
From you

Your data

  • Harvest historyWhat was actually picked, day by day. Entered by hand, or taken from LetsGrow, Source.ag or another system you already use.
  • Your harvest planWeek by week. It competes as a forecast method too.
Fresh every morningYour harvest forecastThe robot’s data, your climate and your harvest, in one forecast.
kg/m²for next week and the week after

What you get

  • Expected yield (kg/m²)for next week and the week after, with a low–high range
  • Where to pick firstthe rows and metres with the most ripe fruit
  • How well it didlast weeks’ forecasts next to what you picked

How it works

The robot sees the fruit. Your greenhouse sets the pace.

Every fruit, every pass: from green to red on a 0–70 scale
Green
Breaker
Turning
Nearly ripe
Red

The robot counts what is coming

Every truss, every fruit, and how far each one has coloured.

Warmth decides when it is ready

A warm week ripens faster than a cold one. The forecast follows your greenhouse temperature, day by day.

It learns from your own harvest

Each week it compares what it said with what you picked, and keeps the method that does best for your crop.

Fruit defects

Only the fruit you can sell

Not every tomato makes it to the box. Scout finds blossom end rot, other fruit defects and botrytis on every pass, and the forecast takes them into account. You also see what share of the yield they cost you, and where they are.

Fruit defects as a share of the yield: all defects 0.2 percent, blossom end rot 0.1 percent, fruit defect 0.3 percent, botrytis 0.0 percent, and 1.14 kg per square metre colouring on the plants
Fruit defects as a share of the yield, with how many finds and in how many places.

HARVEST PLANNING

Not just how much, but where to pick first

The same scans that drive the forecast tell the harvest team which rows, and which metres of each row, have the most ripe fruit at picking height.

Greenhouse map of fruit ripe now per square metre in each row, from the latest scan, with markers where fruit defects are at least twice the usual level
Fruit ripe now per m², row by row from the latest scan. The darker the red, the more to pick; ! marks places where fruit defects are at least twice the usual level.

Where it ends

  • Validated in three greenhouses on three tomato varieties over the 2026 season. Each new variety needs its own calibration.
  • The forecast leans conservative: when it misses, it is usually low, not high.
  • In weeks with a small harvest, read the miss in kg/m², not in percent.

What's next

  • Following each flower to the fruit it becomes, through row lowering, so today's flowers join next month's forecast.
  • Calibration for more tomato varieties.
  • Your own harvest plan as a competing method in every greenhouse.

Same drive, more answers

The same photos do more

No extra hardware and no extra route. The robot is already driving every row, so the same pictures can answer more questions about your greenhouse.

Pests & diseases A map of the places that need your eyes, fresh every morning, with the photo of every find. See how it works → Crop registration Apex growth, internodes and head thickness for every plant in the measured sections, every day. See how it works → Greenhouse diagnostics A daily map of twisted rail sections, with the spots to fix first. See how it works → Labour control Side shoots left on the stem, leaves not removed and missed picking, row by row. See how it works →

Know your harvest two weeks ahead

Scout counts every fruit in your rows and turns it into kg/m² for the next 7 and 14 days. Tell us about your greenhouse and we prepare an offer.

Forecast your harvest with Scout