An illustrative result, not measured crop data. Counts come from visible virtual fruit; volumes come from fitted, capped scene shapes. Mass values below are editable scenarios, not measured biomass.
VISIBLE FRUIT0Unique observed candidates
RED / RIPE APPEARANCE00% of visible fruit
GREEN + PALE00% of visible fruit
TURNING00% of visible fruit
SIZE & UNIFORMITY
A distribution, not just an average.
mL / fruit
—mean modelled volume
— standard deviation
RedGreen / paleTurning
Middle 50%—Range—Total visible-fruit volume—
Shape-based virtual values. A one-sided pass does not establish the complete geometry of a real fruit.
HARVEST PICTURE
Read the colour mix.
Red appearance is a reviewable colour class. Physiological ripeness, hidden fruit and whole-row yield need independent validation.
OBSERVED FOLIAGE SUPPORT—The fraction of sampled virtual leaf surface actually seen in this pass.
GREEN BIOMASS & FRUIT MASS
Make the conversion assumptions visible.
ASSUMED SCENARIO
Observed one-sided leaf area—Virtual visible foliage only
Indicative foliage fresh mass—Area × assumed mass per area
Indicative visible-fruit mass—Modelled volume × assumed density
The starting factors are invented demonstration assumptions. Calibrate them against weighed samples at the facility. This estimates supported foliage and visible fruit only; it does not establish total plant mass, dry biomass or harvest yield.
Enter positive, finite factors to calculate a scenario.
THE VALUE OF A FACILITY VISIT
Turn the row into comparable evidence.
01 / HARVESTA visible fruit inventory
Reviewed red, turning and green counts by gutter and section, with images to revisit uncertain candidates.
02 / UNIFORMITYFruit size and dispersion
Per-fruit size and volume estimates, averages, spread and outliers to compare varieties, sections and visits.
03 / GROWTHCanopy change over time
Visible foliage area and surface coverage across repeat visits, with calibrated mass proxies where evidence supports them.
What makes a field report credible?
Keep row, section and visit identities consistent. Establish scale and camera calibration. Register separate aisle passes before combining them. Review segmentation and fruit identities, record coverage gaps, and compare a sample with manual measurements and weighed material. The simulation illustrates that workflow; it has not run SAM inference or collected sensor data.
Volume and area method details appear with the results.