Industries / Food, farming and forestry / Agriculture and farming / Crop dusting
Mapping and guidance
This work keeps each pass of the aircraft the right distance from the last, turns the spray on and off at field edges, and records where everything went.
5
Milestones achieved, from 1990s to 2009.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1990s
Satellites replace flaggers
Pilots first spaced their passes by eye, then by flaggers, people at the field ends who waved flags and stepped off each swath, and later by mirrors, tethered balloons, paper flags dropped from the wing, puffs of smoke and LORAN-C, a radio navigation system that could not guide swaths narrower than 60 ft (about 18 m). A USDA handbook chapter by Gil Rodriguez and T. J. Roland reported that GPS had already proven itself in US rangeland grasshopper and cotton boll weevil control, with 60 to 100 ft (18 to 30 m) of error, cut to 3 to 6 ft (1 to 2 m) with differential correction (a ground station that broadcasts corrections). By 2007 AOPA, a pilots' association, reported that GPS guidance had "virtually precluded the need for flags".
LessonThe handbook notes that taking ground crews out of the field also took people out from under the spray.
Credit
USDA control programs; guidance makers such as Satloc
2003
Free corrections by satellite
The US Federal Aviation Administration (FAA) put its Wide Area Augmentation System (WAAS) into service on 10 July 2003. WAAS sends GPS receivers corrections worked out by a network of ground stations, so an aircraft no longer needs a base station of its own near the field. By 2007 most operators in northwest Minnesota used WAAS for about 1 m accuracy "at considerable savings in cost and manpower", AOPA reported.
Credit
US Federal Aviation Administration
2008
Brazil files the GPS map
Brazil's agriculture ministry (MAPA) rule IN 2 of 3 January 2008 requires an operational report for every aerial job, kept for at least two years, and when the aircraft carries a recording DGPS, a copy of the application map must be filed with it. The same rule bars aerial pesticide spraying within 500 m of towns and public water intakes, and within 250 m of water sources, isolated homes and livestock.
Credit
Ministério da Agricultura, Pecuária e Abastecimento (MAPA)
2008
A map of sensitive fields
Purdue University's College of Agriculture and the Office of Indiana State Chemist built DriftWatch in 2008, a free, voluntary map on which growers of specialty and organic crops mark their fields so that applicators can check before spraying; beekeepers and seed companies now map hives and seed fields on it too. A non-profit, FieldWatch, now runs it, and lists 28 US states and three Canadian provinces that joined between 2009 (Indiana) and 2026 (Vermont). Virginia's agriculture department hands out its guide for applicators.
Credit
Bernie Engel of Purdue and Steve Smith of Red Gold Tomatoes, whom FieldWatch credits as its founders
2009
Spray rates follow a map
USDA ARS engineers at Stoneville, Mississippi, tested a Satloc M3 guidance unit using WAAS with an AutoCal II flow controller that read a prescription map (a file of zones and the rate each needs) and changed the rate across four zones in flight at about 65 m/s (145 mph). The spray landed on average about 5 m from the intended zone boundary, and the flow controller's error ranged from −1.0% to 2.1%. They called it the first known test of a hydraulically operated variable-rate aerial system.
Credit
Steven J. Thomson, Lowrey A. Smith and J. E. Hanks, USDA ARS
Watch for
Live wind for the pilot
DynaNav Systems and Microair Avionics say that pairing DynaNav's guidance with Microair's T3000 digital air-data and angle-of-attack probe shows the pilot real-time wind speed and direction against the heading and spray path, improving precision and reducing overspray near sensitive crops.
Achieved whenan independent test publishes how closely its in-flight wind readings match a reference instrument, or measures less off-field deposit with it than without.
Claimed by
DynaNav Systems and Microair Avionics, November 2025
Open
01
One format for spray logs
Free software that reads the job logs written by crewed aircraft guidance systems (Satloc, AG-NAV, DynaNav) and by spray drones (DJI Agras, XAG, Hylio), converts each into one open format (flight lines, spray on and off points, rates, swath width, and wind where recorded), and fills a regulator's job record from it. Use Brazil's drone record as the test form: for every job it asks for start and end times, coordinates, crop, hectares, kind of job, product with volume and dose, flight height, temperature, humidity and wind, the aircraft's identity and the nozzle type, plus a map. Release it as a plugin for AgGateway's open ADAPT farm-data framework and as GeoJSON, with test files.
The briefCredit · unclaimed02
Draft the aerial site map
A tool that takes a field boundary and drafts the site map Maine requires before aerial spraying near people: the target area and property lines, flight hazards such as towers and power lines, every place likely to be occupied (homes, schools and the like) within 1,000 ft (about 300 m), and other sensitive areas such as water and crops within 500 ft (about 150 m), all from open data, for the applicator to check and correct.
The briefCredit · unclaimed
Open 01 · the brief
One format for spray logs
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
Free software that reads the job logs written by crewed aircraft guidance systems (Satloc, AG-NAV, DynaNav) and by spray drones (DJI Agras, XAG, Hylio), converts each into one open format (flight lines, spray on and off points, rates, swath width, and wind where recorded), and fills a regulator's job record from it. Use Brazil's drone record as the test form: for every job it asks for start and end times, coordinates, crop, hectares, kind of job, product with volume and dose, flight height, temperature, humidity and wind, the aircraft's identity and the nozzle type, plus a map. Release it as a plugin for AgGateway's open ADAPT farm-data framework and as GeoJSON, with test files.
Why it is still open
Each maker keeps logs in its own system; Satloc's Cloud receives application logs from the aircraft and passes data to billing companies. ADAPT lists plugins for ISO 11783 machine data, cotton gins and a GeoJSON prototype, but none for aircraft. The open tools found each read one maker's files for one purpose: flyfreely-uploader turns DJI Agras T50 logs into CSV files for a fleet platform, and a 2013 JavaScript library reads only field boundaries from Satloc job files. None fills a regulator's record, though Brazil's Portaria 298 has required these drone records, and a monthly report built from them, since 1 October 2021.
Built with what exists
Logs donated by operators; Python with pandas, GeoPandas and GDAL; ADAPT's plugin interface; hourly readings from the nearest public weather station (INMET in Brazil, NOAA's Integrated Surface Database in the US); an AI coding agent to work out undocumented file layouts from sample files and manuals; an AI model that reads long documents to map each record form's fields, with a person checking (see the Note "Seven more flagged").
First step
Ask five aerial operators and five registered drone operators, through their associations, for one season's logs and the records they filed by hand, with farm names removed.
Done looks like
200 jobs from at least three crewed guidance systems and three drone makers converted, with the sprayed area from each converted file within 1% of the acres the maker's own software reports; 30 drone jobs' Brazilian records filled automatically and checked against the records operators filed by hand, with the share of required items filled without typing; converter and test files, with farm names removed, published.
Open 02 · the brief
Draft the aerial site map
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A tool that takes a field boundary and drafts the site map Maine requires before aerial spraying near people: the target area and property lines, flight hazards such as towers and power lines, every place likely to be occupied (homes, schools and the like) within 1,000 ft (about 300 m), and other sensitive areas such as water and crops within 500 ft (about 150 m), all from open data, for the applicator to check and correct.
Why it is still open
Maine's Chapter 22 rule requires whoever contracts an aerial application within 1,000 ft of a "Sensitive Area Likely to Be Occupied" to give the applicator a site map drawn to scale with these features, and every applicator to record sensitive areas within 500 ft and update the record each year. Existing maps cover parts of the job: FieldWatch shows only sites that owners choose to register, and counties such as Colusa, California, publish yearly buffer-zone maps for a few sensitive crops. A search found no tool that drafts the map field by field.
law.cornell.edulaw.cornell.edufieldwatch.comcountyofcolusaca.gov
Built with what exists
OpenStreetMap or Microsoft's open building footprints; the USGS National Hydrography Dataset for water; USDA's Cropland Data Layer for crops; school locations from the US National Center for Education Statistics; the FAA's Digital Obstacle File and OpenStreetMap power lines for hazards; GeoPandas; an AI vision model to flag gardens, hives or play areas on aerial photos for a person to confirm.
First step
Draft maps for ten fields in one county from open data this month and check them by hand against aerial photos.
Done looks like
Maps drafted for 100 randomly chosen fields in one Maine county and compared with maps a person makes from aerial photos and a drive on public roads: at least 95% of the person's sensitive areas found, with every miss and false alarm listed.