Log In Download
Use Case

The Spray Drone Survey Workflow: From Mapping Flight to Prescription

Most spray drone accidents don't happen during spraying — they happen because the survey was skipped. An Agras flying 3 meters above a crop at 10 m/s has no margin for an unmarked power line, a lone tree, or a pole that wasn't on any map.

This is the workflow we use in our own commercial spray operation, built directly into DroneField: a monitoring pilot maps the field, the software processes everything locally in minutes, obstacles get marked on imagery from today — not a years-old basemap — and the spray pilot receives a ready flight plan with boundaries, obstacles, and a variable-rate prescription. Here is how it works, step by step.

Why the Survey Flight Is the Safety System

Satellite basemaps in flight planning apps are often months or years old. New power lines, irrigation pivots, grown trees, and temporary structures simply aren't there. When a spray drone operates at crop height, the survey flight is not an optional agronomy step — it is the primary hazard assessment.

A mapping drone flies the same field first, higher and safer, and produces a high-resolution orthomosaic of the field as it exists today. Every obstacle is visible, every boundary is exact, and the terrain model shows slopes the spray drone must respect. Marking hazards on fresh imagery instead of memory or outdated maps is the single biggest risk reduction available to a spray operation.

The same flight also carries the agronomic payload: if flown with a multispectral drone, it yields NDVI, NDRE or GNDVI maps that drive variable-rate decisions — and even an RGB-only drone gives RGB indices such as VARI and ExG, no multispectral camera needed. One flight, two outputs — safety and prescription. See survey with multispectral drones for which platforms capture what.

Step by Step: Two Pilots, One Project

1. The monitoring pilot flies the survey

A Mavic 3 Multispectral, P4 Multispectral, Matrice 4E or another supported DJI survey drone maps the field in a standard grid mission. This pilot can be a dedicated scout covering many fields per day, well ahead of the spray team.

2. DroneField processes at the field edge

Images import straight from the SD card, and local processing turns them into an orthomosaic, elevation model and NDVI/NDRE maps in minutes — a 20 ha field at 5 cm/px takes under 5 minutes on a Mac mini M4 — with no upload, no waiting and no internet needed. That's the core promise of the software behind the workflow.

3. Boundaries and obstacles are marked

On the fresh orthomosaic, the pilot traces the field boundary — by hand or with the on-device AI field boundary, which follows the field edge from a single click — and marks every hazard as an obstacle or no-spray zone: power lines, poles, trees, buildings, pivots. Because the imagery is from an hour ago, nothing is missing. The AI field boundary documentation covers the tools in detail.

4. The index map becomes a prescription

One click turns the index map into a gridded application map matched to the spray swath, zoned by mean index or by the share of living plants per cell — full rate where pressure is high, reduced or zero where the crop is clean. Optional Luna AI analysis can propose zones and doses with an agronomic explanation, but the operator makes the final call. Choosing the right index for the growth stage matters here; our guide to understanding the maps explains NDVI vs NDRE in practice.

5. The spray pilot receives a ready flight plan

Through DroneField project sharing, the spray pilot gets the complete package — orthomosaic, boundaries with obstacles, and the variable-rate prescription — exported for their machine: DJI Agras (T20 to T100), XAG, or ABZ Innovation, plus Hardi VRA Shapefile, ISOXML V3/V4 and Shapefile for ground rigs. The spray pilot never has to re-survey or re-draw anything.

6. The farmer sees the result

Via a project sharing link, the client opens the field in a web viewer — orthomosaic, index map, boundary and precision map — and downloads the prescription and AI report without installing anything; a one-click PDF survey report is also available. It turns an invoice into evidence.

Why Local Processing Makes This Workflow Possible

The two-pilot workflow only works if the map is ready before the spray drone arrives. Cloud processing breaks it: uploading 30 GB from a field edge on rural mobile data, then queueing for hours, means the survey data is ready tomorrow — and the spray job either waits or flies blind.

DroneField runs the entire pipeline on the monitoring pilot's laptop, offline. The turnaround from landing the survey drone to sending the spray pilot a finished flight plan is minutes. When conditions change fast — disease pressure after rain, a narrow spray window before wind picks up — that turnaround is the difference between a precise application and a missed one.

It also changes the economics: no per-hectare processing fees means the survey flight costs battery time, nothing more. Operators fly more surveys, which means fresher obstacle data and better prescriptions. Pricing is a simple subscription — 699 € + VAT per year per device, with unlimited projects — or a one-off project license for occasional jobs.

Built From Real Spray Operations

This workflow isn't a product manager's diagram — it is how the DroneField team runs its own commercial spraying business. The obstacle-marking tools exist because we've had close calls with unmarked wires. Project sharing exists because our monitoring pilot and spray pilots are different people in different trucks. Project sharing for clients exists because they kept asking to see what they paid for.

If you're evaluating tools for a spraying operation, compare the workflow end to end, not feature checklists: the feature overview shows everything included, and the comparison pages show how DroneField differs from Pix4DFields and DroneDeploy for ag spray work specifically.

Frequently Asked Questions

Why should I fly a survey before every spray mission?

Two reasons: safety and precision. The survey produces a current orthomosaic on which power lines, poles, trees, and boundaries are marked from today's imagery — most spray drone accidents trace back to obstacles missing from outdated basemaps. The same flight yields vegetation index maps that turn blanket spraying into variable-rate application.

Can the survey pilot and spray pilot be different people?

Yes — DroneField is designed for exactly that. The monitoring pilot processes the survey and shares the project; the spray pilot receives a ready flight plan with boundaries, obstacles, and the prescription map, exported for DJI Agras, XAG, or ABZ Innovation equipment.

What is a drone prescription map?

A prescription (variable-rate application) map divides the field into zones with different application rates, generated from vegetation index maps like NDVI, NDRE or RGB indices such as VARI. DroneField exports prescriptions for DJI Agras, XAG and ABZ Innovation spray drones, plus Hardi VRA Shapefile, ISOXML V3/V4 and Shapefile, so the sprayer varies its rate automatically.

How fast can I go from survey landing to a spray-ready flight plan?

Minutes. DroneField processes locally — a 20 ha field at 5 cm/px reaches orthomosaic and NDVI in under 5 minutes on a Mac mini M4 — and boundary drawing, obstacle marking and prescription export are done in the same sitting at the field edge, with no internet connection required.

Which drones fit this workflow?

Survey side: DJI Mavic 3 Multispectral, Phantom 4 Multispectral, Phantom 4 RTK, Mavic 3 Enterprise, Matrice 4E/4T, and in beta the Matrice 5TD and M300/M350 RTK with Zenmuse P1 — DJI drones only. Spray side: DJI Agras T20 to T100, XAG P40/P100/V40 and ABZ Innovation L-series, plus Hardi VRA Shapefile and ISOXML V3/V4 export for ground sprayers.

Have more questions? Contact our team

Ready to turn drone data into field decisions?

Download DroneField and start analyzing your fields with one free project.