The question, in the buyer's words
I run a 1,600-acre tree nursery where field crews walk every row counting seedlings lost to frost; could drones count the empty spots and flag frost-killed rows before the spring replant?
Key takeaways
- A drone can flag suspect rows, so crews walk fewer rows before the replant order is written.
- Frost damage detection from drones has been tested on crops, not on nursery seedlings.
- Seedling counting studies report strong accuracy on trial plots, not on whole farms.
- Flights beyond visual line of sight can need an FAA waiver, and waivers can take months.
- Check every drone count against ground walks before it drives a replant order.
Can a drone count empty seedling spots in tree nursery rows?
When frost hits a tree nursery, the losses show up row by row, not on a screen. Most operations find them by sending crews down every row and tallying the empty spots, and the replant order waits on that tally. A drone flying over the beds can produce a map of gaps and weak patches in a single pass, so the crew can start where the losses are instead of at the end of the first row.
The seedling counting research is encouraging but narrow. A Chinese study of Korean pine seedlings in a nursery garden used drone imagery with a YOLOv4-based method, aimed at the large errors that standard sampling produces in dense nurseries (Traitement du Signal, 2023). A Finnish study of spruce seedlings (Forests, 2023) reported a mean average precision of 84% and a detection accuracy of 97.86%, and its authors called consumer drones a cost-effective way to inventory seedlings. A 2024 Chinese study trained a counting model on 1,520 drone images of nursery stock and also reported high counting accuracy.
Read those figures carefully. Mean average precision and detection accuracy describe how well a model matches seedlings in test images, not how close a field total comes to the true count across 1,600 acres. A conifer stocking survey (Forests, 2020) reported 8% omission and 6% commission errors for its best height-estimation approach. Those two error types push a count in opposite directions, so one accuracy figure can hide the error that matters most to your replant order.
How well do drones pick out frost damage in seedling rows?
Frost detection from the air has been tested mostly on crops. In a 2024 open-field lettuce study, a model estimating a frost damage index from drone imagery improved its accuracy by 3% to 14% when RGB and multispectral imagery were combined (Frontiers in Plant Science, 2024). The authors also said multispectral cameras are costly and offer low resolution with limited texture detail, which matters when damage shows up as small changes in leaf color.
A 2019 maize study in India flew multispectral drone imagery before and after a frost event and classified the damaged leaves with random forest and support vector machines. Frost-treated leaves showed reflectance changes between 500 to 680 nm and 730 to 740 nm (ISPRS Archives, 2019). That is useful evidence that frost leaves a spectral signature, but it comes from maize, not from a young conifer seedling in a bed.
We found no study of frost damage in forestry nurseries, so treat any drone frost flag as a hypothesis to verify on the ground. Frost injury can take time to show in the foliage. Schedule the flight once symptoms are visible in your test rows, and keep a few marked control rows that the crew checks by hand, so you can see what the imagery missed.
How big is the job, and who walks the rows today?
Outdoor nursery work is concentrated in fewer operations than it was. The 2017 Census of Agriculture counted 15,092 outdoor production nurseries, down from 19,184 in 2012, and outdoor nursery acreage fell to 337,855 acres from 404,382 over the same five years (a trade publication reporting the census). The 2019 census counted 6,458 nursery stock operations, down nearly 2,000 since 2009 (Agricultural Marketing Resource Center). Those are sector totals, not your farm, but they show how the base of operations has shrunk.
Labor is the other pressure. A third-party occupational data site, citing the Bureau of Labor Statistics, put crop, nursery and greenhouse farmworkers and laborers at about 504.8 thousand employed in 2024, projected to about 488.1 thousand by 2034. US turnover figures for nurseries were not retrieved in our search. A 2024 Canadian Horticultural Council factsheet reported 18% turnover in nursery, floriculture and ornamentals for 2022, against 13% for farm labor overall. That is Canadian data, so read it as a warning sign rather than a US benchmark.
A 1,600-acre site makes the ground walk a large crew commitment. Each week spent walking is a week the replant order waits, and the crew is pulled from other work the season also needs. The job is not hard to explain. It is slow to finish, and that is where a drone pass can help.
Does a drone flight over 1,600 acres need an FAA waiver?
It depends on how the aircraft flies. Under FAA Part 107, flying beyond visual line of sight for agricultural work needs a case-by-case waiver, and waivers can take months, according to an agricultural drone industry blog. Flights that stay within visual line of sight need no beyond-visual-line-of-sight waiver, so a pilot on a single block can often start sooner than a whole-farm plan that depends on a waiver.
The rules are moving. The FAA proposed Part 108 on August 7, 2025, which would allow drones up to 1,320 pounds to fly beyond visual line of sight, compared with 55 pounds under Part 107. Finalization was expected in 2026, and we could not confirm in our search whether it has been finalized, so check the FAA's own status before you write a plan around it. A June 2025 executive order also directs faster processing of beyond-visual-line-of-sight waivers.
A waiver can be narrow. One Part 107 beyond-visual-line-of-sight waiver granted to a Florida ranch allows flight up to 1,000 feet with visual observers and is valid for up to two years, according to a press release on a newswire service. Ask any provider to show the scope of its waiver in writing, including altitude, observer numbers and the area it covers, before you accept a flight plan.
What does a drone flight day look like on a nursery?
A workable flight day starts before sunrise, when light and wind are steady enough for clean imagery. The pilot checks the flight plan against the block map, confirms the excluded areas such as the loading yard, and marks the control rows that the crew will walk by hand.
The drone then flies a fixed pattern over the test block at a set height and camera setting, so each image can be compared with the next flight. The images are stitched into a map that flags gaps, weak patches and low-ground frost pockets. The output should be a report the crew can use in the field, not a folder of raw pictures.
The crew then walks the flagged blocks and the control rows, and their counts go back into the calibration. The replant order is written from the combined result, with the drone flags and the ground counts side by side.
How do you check a drone count against the ground?
Calibration is what turns a drone map into a count you can order seedlings from. Before the replant window, pick rows that span your frost exposure, from low ground to open beds, and count every seedling by hand in them. Compare the drone result row by row, not only as a farm total, because farm averages can hide the rows that were missed.
The conifer stocking result shows why this matters. An 8% omission rate and a 6% commission rate mean the method can both miss true gaps and misjudge occupied spots, and each error moves the replant order in a different direction (Forests, 2020). Those rates came from one survey with its own method, so your own rates will likely be other numbers. Measure them on your species and bed layout.
Turn the results into a written rule for the whole block, such as which flagged rows always get a crew walk regardless of the drone output. Keep the rule and the calibration numbers on file, so the replant decision can be traced later if a block underperforms.
Where does a drone stop and the crew start?
A drone sees canopy and ground from above. It does not see roots, stem damage below the surface or whether a stressed seedling will recover, so the replant decision still needs a person on foot. Frost-killed and frost-stressed plants can look alike in an image, and that gap often decides whether to replant now or wait.
Crews also handle what the aircraft cannot. They confirm species and grade on a flagged row, check bed labels, and decide whether a patchy row needs full replanting or infill. A drone does not replace that judgment. It narrows the walk from every row to the rows that need judgment.
Keep the roles clear on paper: the drone finds where to look, and the crew decides what a gap means.
How do you start a drone pilot before the spring replant?
Start small and work back from the replant calendar. Choose one block that had frost exposure, fly it within visual line of sight unless your waiver covers more, and use the crew walk to score the drone flags. If the flags hold up against the ground counts, extend the same method block by block before the replant order is due.
We are a full-service integrator for operations like this. We run a free site assessment, map the block layout and airspace, and connect the drone output to the replant planning your team already uses. We choose the aircraft for the site across manufacturers, then lease it through financing, with a fixed lease payment quoted after the site assessment. Deployment, crew training and maintenance under our service plan come in the same package, backed by regional service engineers in every state we service.
Ask for the pilot scope, the calibration plan and the flight rules in writing, and judge the pilot by how well the flags matched the crew walk. A drone is a faster first look at the blocks, not a replacement for walking the rows that decide the replant.
Frequently asked questions
Want a drone at your tree nursery seedling rows?
Tell us about your site. We will recommend the robot, quote the lease, and keep it serviced.
Sources
- Traitement du Signal, 2023 (Korean pine seedling drone counting)
- USDA NASS, 2024 Census of Horticultural Specialties release
- USDA NASS, 2024 Horticultural Specialties highlights report
- Agricultural Marketing Resource Center, nursery trees
- Trade publication report on the USDA Census of Agriculture
- Frontiers in Plant Science, 2024 (lettuce frost damage index)
- ISPRS Archives, 2019 (maize frost reflectance)
- Agricultural drone industry blog on FAA Part 107 and Part 108