Creating Prescriptions Based on PlantAI Outputs

The PlantAI-based prescription workflow allows users to generate variable-rate application maps using detailed plant-level data.

Unlike vegetation index–based prescriptions that rely on canopy reflectance, this method uses direct detections of individual plants — making it especially useful for high-value or row crops such as lettuce, sugar beet, or vegetables.

By leveraging PlantAI results such as plant size, density, or weed detection, growers can create precise prescriptions for fertilizer adjustment, replanting, or spot spraying.

Step 1. Start a New Prescription

Open the Prescription Tool and click New to create a new prescription file.

When selecting the data source, choose PlantAI Detection.

In the drop-down list, select the detection run (e.g., “Detection 1”) that corresponds to your desired dataset — for example, a run where plants or weeds were detected.

Step 2. Configure the Grid

Define the grid used to segment the field into manageable rate cells.

This grid determines the spatial resolution of the prescription and should align with the machinery’s working pattern.

You can adjust:

• Rotation to match the driving direction of the equipment.
• Working width to correspond to the implement’s actual coverage (e.g., 3.5 m for a precision fertilizer spreader).
• Cell height to set how frequently rate adjustments occur (e.g., every 5 m).
• Position by dragging the grid so it aligns visually with tramlines or crop rows.

A correctly aligned grid ensures that prescription changes occur in sync with the machinery’s movement, minimizing overlap or missed areas.

Step 3. Choose the Metric

PlantAI provides multiple metrics that can be used to generate prescription zones:

After choosing the metric (for example, Average Plant Size), click Create Prescription.

The system will calculate the chosen metric for each grid cell and visualize it as a classified map.

Step 4. Define Classes and Thresholds

Once the initial prescription map is generated, the data can be divided into classes — for example, small, medium, and large plants.

By default, three classes are created, but this can be customized.

You can:

• Add or remove classes depending on desired rate complexity.
• Adjust thresholds manually using sliders to define the size or density limits between categories.
• Rename classes for clarity (e.g., “Low Growth,” “Normal,” “High Growth”).

This classification determines how input rates (e.g., fertilizer) will be distributed across the field.

Step 5. Assign Rates

For each class, specify the rate corresponding to the crop’s condition.

A typical approach might be:

• Small plants → higher fertilizer rate to promote catch-up growth.
• Medium plants → standard rate for maintenance.
• Large plants → lower or zero rate to avoid over-application.

As you assign rates, a summary of the total and average application amounts is displayed for review.

This helps confirm that total nutrient use aligns with your agronomic plan and available resources.

Step 6. Save and Export

Once satisfied with the configuration, click Save to store the prescription.

To download, navigate to the Export tab and select Prescription.

Available export formats include:

• Generic Shapefile (.shp)
• John Deere-compatible format
• DJI Agras drones
• Hardi GeoSelect

Best Practices

• Ensure that all detections in the selected PlantAI run are verified and correspond to the correct crop type.
• For size-based prescriptions, perform detection at a consistent growth stage where plant diameter represents relative vigor accurately.
• Align the grid precisely with machinery driving direction and working width.
• Avoid excessively small grid cells, which may lead to noisy data or overly complex prescription files.

By combining plant-level data with automated rate mapping, PlantAI-based prescriptions enable highly targeted, data-driven management for specialty and high-value crops.