Log In Download
Learn

How to Capture Good Drone Monitoring Imagery: Flight Settings, Weather and SD Cards

A vegetation index map is only as good as the images behind it. The processing software can only work with what the drone photographed — nothing can make up for a missing strip, a blurred frame or a cloud shadow baked into the data. Most disappointing NDVI maps are not a software problem; they were decided on the field, before takeoff.

This guide covers the five things that determine map quality for multispectral and RGB surveys: flight settings, weather, the SD card, processing habits and a pre-flight safety routine. It is written for beginners, but the numbers are the ones experienced operators use. All of it, with a printable checklist, is in our free handbook <em>How to capture great monitoring imagery</em> — available in the <a href="/en/knowledge-base">DroneField Knowledge Base</a> after registration.

1. Flight settings: overlap, gimbal, altitude, RTK

Overlap: 75–80% front, 65–75% side

The software stitches images together using the areas they share, like a jigsaw puzzle. Too little overlap and the map ends up with gaps, distortions or misaligned strips. Use 75–80% front (along-track) overlap and 65–75% side (cross-track) overlap. Uniform crops such as a closed cereal canopy have few recognisable features, so go to the higher end of both ranges there.

Gimbal: nadir, –90°

Point the camera straight down. Oblique images distort both the geometry and the vegetation index, because the plants are seen from the side at a different angle — and the index is a reflectance measurement, not a photo.

Altitude: let the goal decide

Over farmland, 90–115 m AGL (height above ground level) is the usual range. Flying higher is faster, needs fewer images and covers more area per battery; flying lower gives a more detailed map but a slower job and more data. A prescription map for a spray drone does not need full resolution — management zones are metre-scale anyway, so the upper end of the range is plenty. Stand counts, weed patches and detailed analysis need the lower end.

Pattern, terrain follow, RTK

Fly back-and-forth parallel lines along the long side of the field: fewer turns, more consistent images. On hilly ground enable terrain follow so the height above ground, and with it the resolution, stays constant. And fly with RTK: NTRIP services are now free or cheap, images align more accurately, and later surveys stack on top of each other, which is what lets you track change in a field over the season. If you are new to the processing side, our guide to how drone field mapping works explains what happens to the images next.

2. Weather: the measurement is only as stable as the light

A multispectral camera measures light reflected by the plants. If the light or the plants move or change during the survey, the measurement becomes inaccurate. That is why weather matters at least as much as the flight settings.

Wind: no more than 5–7 m/s

In strong wind the drone wobbles, images can blur and the battery drains faster. Wind also moves the plants: the same leaf sits in a different spot in two consecutive images, which hurts alignment.

Clouds: clear, or uniformly overcast

Best is a cloudless sky, or a continuous, even cloud cover that shades the whole area the same way. Do not fly under broken cloud: moving cloud shadows leave patches on the map that the vegetation index will show as weaker crop — and you will not be able to tell them from real stress later.

Sun angle: around solar noon

When the sun is highest, shadows are shortest and the light is strongest. Early morning and late afternoon bring long shadows and weak light of shifting colour that degrade the measurement.

Dew, fog, after rain

Wet leaf surfaces reflect like a mirror and distort reflected-light measurements. Wait until the dew has dried; do not fly in fog or drizzle. Cold also reduces battery capacity noticeably, so pre-warm the cells and plan shorter flights.

One habit pays off more than any other: if you survey the same field regularly, do it at a similar time of day and in similar weather. That makes maps comparable, so what you see is real change in the crop, not a difference in lighting.

3. The SD card: write speed, not the big number on the box

At every capture point a multispectral camera saves several images at once — one RGB frame plus one per band — while the drone keeps flying. If the card cannot write fast enough, frames are skipped and you get a hole in the map.

The markings that matter are the U and V classes, which guarantee a minimum sustained write speed: U3 means at least 30 MB/s, V30 the same, V60 and V90 more. Plain Class 10 (10 MB/s) is not enough any more. The large number printed on the packaging, such as "up to 170 MB/s", is usually the read speed — ignore it for capture. UHS-I is sufficient, because most drones run UHS-II cards at UHS-I speed anyway, and the A1/A2 app-performance rating is irrelevant here.

For size, 64–128 GB is usually plenty for a working day, but multispectral surveys fill storage faster than you expect, so always carry a spare and check your drone manufacturer's supported sizes. Buy branded cards from a reliable source (counterfeits are far slower than their markings), format in the drone before every major job but only after the previous data is backed up, do not fill the card to the brim, remove it only with the drone off, and back up after every field before clearing it.

4. Processing habits that save hours

Copy the images to your computer — ideally an internal SSD — and work from there. If you work from an external SSD, use a data cable rated at least 10 Gbit/s. Two cables can look identical (both USB-C) while one copies seventy times slower: most phone charging cables are USB 2.0, where a 10 GB survey takes about five minutes to copy instead of eleven seconds over USB 3.2 Gen 2. Look for a 10, 20 or 40 Gbps rating or the Thunderbolt logo on the cable, and keep the cable that came with your SSD labelled and separate. Copy speed is set by the slowest link in the chain — card, reader, cable, port, disk — so copy once and work locally.

Organise surveys into folders by field and date, and never rename or cherry-pick files: multispectral band images belong together, and one missing file degrades the result.

Match the processing resolution to the goal. A spray-drone prescription map is best processed at reduced resolution — faster and smaller, and the zones are metre-scale anyway. Crop assessment and vegetation indices work well at medium to native resolution. Only stand counts and detailed weed analysis need the native high-quality setting, because individual plants and weed patches only separate at high resolution.

After processing, review the orthomosaic for missing corners, misaligned strips or blurry areas, look for patches that suggest cloud shadow, and keep the original images so you can reprocess with different settings at any time. Step-by-step instructions for every DroneField feature are in the public user manual.

5. Safety: a pre-flight checklist you actually use

Go through a short list before every operation — print it and keep it with the equipment. The site and weather part covers a stable sky and light wind, no rain or thunderstorm expected within an hour of landing, airspace and temporary restrictions checked, permits in place, only involved and briefed persons in the flight area, the takeoff site secured and surrounding obstacles known.

The pilot and drone part covers a valid and appropriate pilot qualification, insurance, charged batteries with voltage and temperature OK and a spare at hand, an undamaged airframe, propellers and camera with clean lenses, a charged controller, a solid drone–controller link with continuous camera feed, GNSS working and RTK fixed, an SD card inserted with enough free space, failsafe settings checked with the return-to-home altitude set above surrounding obstacles, and the flight plan loaded with the parameters set for the intended flight.

The free handbook contains this checklist as a printable one-page form with date, location, pilot, drone serial number and signature fields. It does not replace documents required by your regional regulations, but it catches the mistakes that most often end a survey early.

Get the free handbook

Everything above — with the overlap diagram, the weather table, the SD card marking guide, the cable speed comparison and the printable checklist — is in How to capture great monitoring imagery, one of the two handbooks that are free after registering a DroneField account. It is available in nine languages in the Knowledge Base of the customer portal, alongside the subscriber handbooks on vegetation indices and crop-specific monitoring. See all handbooks or register and download it now.

Frequently Asked Questions

What overlap should I use for multispectral drone mapping?

75–80% front overlap and 65–75% side overlap. Uniform crops with few recognisable features, such as a closed cereal canopy, need the higher end of both ranges so the images can still be aligned.

What altitude is best for NDVI mapping of farmland?

90–115 m above ground level is the usual range. Use the upper end for spray-drone prescription maps, where zones are metre-scale anyway, and the lower end for stand counts, weed patches and detailed analysis.

Can I fly a multispectral survey on a cloudy day?

Yes, if the cloud cover is continuous and even for the whole survey. Avoid broken, moving cloud: the shadows appear on the map as patches of apparently weaker crop that cannot be told apart from real stress later.

Which SD card do I need for a multispectral drone?

A branded card rated at least U3 / V30, typically 64–128 GB, from a reliable source. Write speed is what matters for capture; the large 'up to' number on the box is usually the read speed.

Do I need RTK for crop monitoring?

It is strongly recommended. RTK aligns images more accurately and lets later surveys stack precisely on earlier ones, which is what makes change over the season measurable. NTRIP correction services are now free or inexpensive in most regions.

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.