Why it matters
Without AI, botrytis risks rise due to missed early signs, humid microclimates, and inconsistent records, leading to losses in blooms, quality, and certified status.
From design to safety
Design
Define sensors and data streams
Assess potential data sources: petal spotting, leaf wetness, humidity, debris accumulation, and historical fungicide logs to shape AI capability and early risk indicators. Map data governance, sensor redundancy, and data retention rules for audit readiness and scalable monitoring across multiple tunnels and seasons.
- leaf wetness sensor
- humidity sensor module
- canopy camera array
Build
Install data capture and routing
Install sensors, cameras, and loggers; establish wired or wireless networks, calibration routines, data pipelines to a central AI model for real-time assessment, and redundancy planning. Validate data quality through ground-truth scouting feedback and create rollback plans for sensor outages.
- leaf wetness sensor
- humidity sensor module
- canopy camera array
- ventilation fan control
Operation
Real-time risk scoring and alerts
AI ingests petal spotting data, humidity, and debris images to score Botrytis risk; triggers scouting routes, fungicide timing prompts, and packing plan updates for bloom integrity. Continuously refine thresholds using season-long trend analyses and operator feedback.
- canopy camera array
- ventilation fan control
- fungicide storage cabinet
Maintenance
Record integrity and data health
Regularly back up logs, verify sensor calibrations, retain fungicide application records, audit data consistency, and append postharvest hydration data and debris sanitation checks for traceability. Implement periodic firmware reviews and secure archival of legacy data.
- fungicide storage cabinet
- humidity sensor module
- leaf wetness sensor
Safety
Audit trail and regulatory checks
Maintain verifiable records for chemical handling, safety training, and GLOBALG.A.P./MPS compliance verifications; ensure approvals by qualified personnel in audits and re-certifications. Prepare incident logs and corrective action reports on any data anomaly.
- fungicide storage cabinet
- ventilation fan control
- canopy camera array
Part by part
- 1
leaf wetness sensor
- Watch for
- Inaccurate surface moisture readings or sensor fouling on leaf surfaces; failure to detect rapid dew formation.
- AI action
- AI reads leaf moisture patterns and flags rising infection risk with regional heatwave adjustments.
- 2
humidity sensor module
- Watch for
- Drift or calibration loss causing mistimed irrigation and stagnant microclimates; sensor shadowing near dense canopy.
- AI action
- AI triggers alert for recalibration and adjusts humidity thresholds; logs drift corrections.
- 3
canopy camera array
- Watch for
- Subtle petal spotting or color changes indicating early Botrytis; misclassification between normal bloom aging and infection.
- AI action
- AI marks suspected hotspots and generates scouting routes with confidence scores.
- 4
ventilation fan control
- Watch for
- Stagnant air, insufficient air exchange, high humidity pockets; uneven cooling across tunnel sections.
- AI action
- AI opens/closes vents and adjusts fan speed for airflow balance and distribution.
- 5
fungicide storage cabinet
- Watch for
- Expired or mislabeled product, improper access control; missing batch records.
- AI action
- AI flags stock nearing expiry and prompts rotors to restrict access until verified; enforces lot traceability.
- 6
cooler and case log
- Watch for
- Inconsistent temperature or humidity logs, risk of degraded stems; door seals failing.
- AI action
- AI prompts daily cooler checks, records adjustments, and risk-based relocation guidance for high-risk lots.
Worked example
Detecting Botrytis onset in a 12-acre cut flower operation
Illustrative scenario with sample data.
A mid-season protocol triggers when humidity rises and canopy images show minor petal spotting; the system flags a regional cluster for proactive intervention before symptoms expand.
Signal in
area_acres:12, tunnels:5, sensor_readings: [{type:humidity, value:78%},{type:leafWetness, value:moderate},{type:imageQuality, value:OK}], fungicide_records: [{date:2026-10-28, product:STF-Variant, lot:LV-4821}]
System out
Alert: Botrytis risk elevated in tunnels 2, 3, and 5; recommended scouting route across blocks A to E; suggested fungicide window: Day 3 post-flag.
Human sign-off: Head grower reviews the alert, approves scouting plan and fungicide timing, signs off on the action report, and updates the risk dashboard for senior management.
How to roll it out
Phase 1
Weeks 1 to 4: Baseline
Install sensors, camera array, and loggers; establish data routing and initial risk scoring across all tunnels.
Phase 2
Weeks 5 to 8: Pilot scoring
Run pilot on three tunnel blocks; validate alerts, adjust thresholds, and refine audit trails and data quality checks.
Phase 3
Weeks 9 to 12: Scale and formalize
Extend system across all tunnels; implement standardized response playbooks, training modules, and certification-ready records.
Where people stay in charge
- All Botrytis risk decisions above low threshold remain with the head grower or designated crop-health lead.
- Live pesticide handling and application decisions must be approved by a licensed applicator or certification manager.
- Critical records such as fungicide use and audit logs must be signed off by the farm owner or responsible safety officer.
Questions owners and crews ask
What signals does AI botrytis detection read on the canopy?
It reads petal spotting patterns, leaf wetness, and real-time canopy images to assign risk levels and trigger scouting or interventions. The system weighs color shifts, texture changes, and browning edge symptoms to prevent missed early signs.
How does the system handle humidity and debris data?
Humidity readings feed microclimate models while debris imagery informs sanitation checks for infection risk; debris classification differentiates plant matter from dust, aiding hygiene protocols.
Who approves fungicide actions suggested by the AI?
The head grower or licensed applicator reviews and signs off on fungicide timing and product choice; they also confirm resistance management and label compliance before application.
What records must be kept for certification?
Fungicide logs, application dates, batch numbers, safety trainings, and audit-ready data trails; include spray diaries, calibration certificates, and postharvest quality checks for GLOBALG.A.P./MPS audits.