Industry field guide
Planting pass data for evaluating AI field maps
A line on a map proves a machine moved.
- 2022 NAICS
- 115112
- Practice pack
- Example and answer key (JSON)
- Updated
What this guide tests
Which distance is confirmed active, which path is repositioning, and what cannot be counted as verified unique planted area?
See the specimen and fields · Work through the fictional records
Task and source context
A line on a map proves a machine moved. A planter’s state record can help show which part of that path actually represents planting. NRCS publishes precision-farming guidance, and Blue River Technology describes agricultural applications that use computer vision. A service contractor can propose a narrower test of evidence alignment using anonymized or simulated passes. NRCS precision farming guidance · Blue River Technology applications
Sketch a demonstration field and generate short parallel paths. Mark one pass with an active planting signal, one with the tool raised, and one where two sections overlap and a state transition happens late. Match each trace to a verified operator note. An evaluator can ask whether the model credits only confirmed planted distance and identifies uncertain intervals. Use a separate block for testing.
Planting maps and variety assignments may belong to the client farm. Remove real coordinates, names, seed rates, and equipment identifiers. Confirm data export and reuse authority before discussing an actual season.
Industry research
The U.S. Census Bureau’s industry definitions establish which activities these codes cover. The workflow sources in this guide describe the task or record type. Fiund’s pilot below uses fictional material to test a specific decision within that scope.
115112 · Soil Preparation, Planting, and Cultivating
This U.S. industry comprises establishments primarily engaged in performing a soil preparation activity or crop production service, such as plowing, fertilizing, seed bed preparation, planting, cultivating, and crop protecting services.
Build a pilot sample
Set up this pilot
Draw a simulated planting route with overlaps and an active-state transition. Measure confirmed active distance separately from repositioning and independently verified unique area.
Prepare the test packets
- Start with the worked case. Use the fictional record types below as a guide to the source files you need. Give each packet a stable ID and save its source versions with it.
- Change one fact. Make one packet with enough evidence to answer the question and one where the evidence points to a different answer. Keep the other fields fixed so a reviewer can tell what changed.
- Remove the decisive evidence. Give the reviewer a third packet with a plausible near-match in its place. Have them write what can still be established and what evidence would resolve the rest.
Fields the reviewer needs
- Pass and path segment
- Identifies spatial intervals.
- Tool state and transition uncertainty
- Separates activity from travel.
- Overlap, width and operator source
- Defines what further evidence area calculation needs.
Use Plan a farm, forest or fishery AI data sample to turn the specimen into a brief with a defined task, source and review owner.
Fictional exercise
A machine path needs tool-state evidence before it counts as planted distance
Abstract parallel paths and original state annotations form an offline contractor-data exercise. There are no real coordinates, seed rates, machine settings or operating instructions.
| Record ID | Source | What it records |
|---|---|---|
| PPF1 | Pass A | Path length 100 practice meters; planting-state signal active throughout; matching fictional operator note confirms the interval. |
| PPF2 | Pass B | Path length 100 meters; tool raised throughout; operator note says repositioning. |
| PPF3 | Pass C | Path length 100 meters overlaps A for its first 20. Active-state transition timestamp uncertain over the first 30 meters; final 70 confirmed active. No width or area denominator supplied. |
Question for the model. Which distance is confirmed active, which path is repositioning, and what cannot be counted as verified unique planted area?
Show the reference answer
PPF1 supports 100 practice meters of active-state travel on A. PPF2 is 100 meters of repositioning with the tool raised. PPF3 confirms only its final 70 meters active and leaves the first 30 uncertain, including overlap with A. These records do not establish unique planted area because width and adequate overlap treatment are absent. Report supported distance by pass rather than credit every mapped line as planting or sum overlap into area.
How to check the result
Ask for the answer, the record IDs that support it and any fact the evidence leaves unresolved. Have a reviewer compare the response against the reference before deciding whether the sample is ready to expand.
Errors to flag in this case
- Treating all machine travel as active planting.
- Summing overlapping paths into unique area without tool-state timing and width evidence.
Change the inputs and run it again
Provide C's exact activation at meter 30 and an explicit width while leaving overlap geometry incomplete. Require the remaining area gap to stay visible.
Score the decision and the evidence separately
Save the reviewer’s expected answer, supporting record IDs and unresolved fields in a separate answer key. Score the model on the decision, its citations and its handling of the deliberate gap. A correct answer citing a different case counts as an evidence error.
Before expanding collection, have a second reviewer run the same checks on the held-out packet. Keep related versions, views and exports of the same case together so they stay on one side of the training and evaluation split.
Check permission and scope
Use simulated paths. Exclude client maps, variety assignments and equipment IDs; verify client and platform export authority before real collection.
Use authorized notes or recreated records. Keep crew rosters, customer identities and private field details outside the sample.
Describe an authorized archive through Fiund’s existing-data assessment, or describe a new capture capability through the collection brief. The contributor agreement covers the terms for participation.
Sources and further reading
Resources for this workflow
Company and program reviews
Use these reviews to compare sourcing, capture or evaluation approaches. Confirm the applicable program, accepted sources and terms with the provider.
Compare the approach
truelabel review for teams collecting physical AI data
Compare how a capture team specifies devices, processing and quality review for an outdoor recording pilot.
Compare the approach
Graviti review for workplace video and multimodal training data
Compare a catalog with commissioned multimodal collection when a particular crop, setting or viewpoint matters.
Related industry examples
Industry guide · NAICS 115115
Farm crew work-order records for AI evaluation
A crew leader can plan a job at one site and finish it at another.
Industry guide · NAICS 115116
Farm management acreage revision data for AI
A revised field boundary can change area calculations without changing what was planted.
Industry guide · NAICS 115111
Cotton gin bale lineage data for AI traceability
A gin is in a different position from a grower.