Palm Tree Health Records with AI Validation
Inspectors record the health of palm trees in the field, with exact coordinates and the cause of disease — and an AI model checks whether each tree is really sick.
The problem
Palm tree diseases spread across many farms and locations. To act early, you need to know which trees are sick, where exactly they are, and why. Field inspectors can record this — but two inspectors may judge the same tree differently, and a wrong record sends treatment to the wrong place.
What I did
- Built an inspector app to record each palm tree’s health in the field, with its coordinates and the cause of disease.
- Built the admin console where the team reviews all records across locations.
- Connected an AI model that checks each record and says whether the tree is really sick.
- Added validation that compares the inspector’s values with the AI model’s, and flags the cases where they differ.
Architecture
-
Field
- Inspector app
- Coordinates & cause of disease
-
Platform
- Health records
- AI model check
- Inspector vs AI validation (key component)
-
Office
- Admin console
The hard part: When the inspector and the AI disagree, the record is not trusted yet.
Neither side is always right. An inspector can misjudge a tree, and a model can misread it too. So every record carries both opinions: what the inspector saw, and what the AI model says.
The system compares the two. Where they match, the record can be trusted. Where they differ, the difference is shown clearly so the team can review it — and over time, these cases show which inspectors and which conditions need a closer look.
Results
- GPS exact location for every tree
- 2 apps: inspector app and admin console
- AI second check on every record