Drone Sensors for Crop Health Surveys: What Each One Does

A drone crop survey is only as good as the sensor under it. The main drone sensors for crop health surveys — RGB cameras, multispectral sensors (which produce NDVI), thermal cameras, LiDAR, and the SLAM technique that keeps a drone flying where GPS is weak — each measure something different. This is a plain-language tour of what each one is and the purpose it serves.

You rarely need all of them. Knowing what each sees is how you pick the right tool, or the right combination, for the job in front of you.

An RGB camera is the ordinary colour camera on almost every drone, and it's where most surveys start. It captures visible red, green, and blue light, and its overlapping photos stitch into a sharp orthomosaic you can measure: field area, plant and tree counts, planting gaps, roads, and storm damage. With enough overlap it also builds a basic 3D surface model through photogrammetry.

What it can't do is see inside the plant. Early stress — before leaves visibly yellow — looks the same as healthy crop to an RGB sensor. It maps the farm; it doesn't diagnose it.

A multispectral sensor is the workhorse of crop-health surveying. It records bands beyond visible light — most importantly near-infrared and red-edge — which healthy leaves reflect strongly. Comparing those bands produces vegetation indices like NDVI and NDRE that score plant vigour pixel by pixel, often catching stress days before it's visible to the eye.

NDVI itself isn't a sensor; it's the index a multispectral sensor's bands are turned into. For the full method, see our guide to NDVI drone mapping.

A thermal camera measures temperature instead of colour, and on a farm that mostly means water. Plants cool themselves by transpiring, so a stressed or water-short plant runs warmer than a healthy, well-watered one. A thermal map exposes dry corners, blocked irrigation, and patches under stress, and it can flag some diseases early through the heat signature of damaged tissue. It's a strong complement to multispectral: one reads vigour, the other reads water.

A hyperspectral sensor is multispectral taken to an extreme — instead of a handful of bands, it records dozens or hundreds of narrow ones. That fine detail can separate specific diseases, nutrient deficiencies, or even crop varieties that broader sensors blur together. The cost is real: the sensors are expensive, the data is heavy to process, and it's mostly used in research and large commercial operations rather than everyday survey work.

LiDAR measures distance with laser pulses, building a dense 3D point cloud of the crop and the ground. Where cameras see a flat surface, LiDAR sees structure — canopy height, density, gaps, and biomass — and its pulses slip through canopy gaps to map the terrain underneath. That makes it the tool for tree and plantation crops where height and the layer beneath the canopy matter. See our guide to LiDAR for canopy health.

SLAM — Simultaneous Localization and Mapping — isn't a payload sensor like the others; it's a technique a drone uses to know where it is and build a map at the same time, using cameras, LiDAR, and an inertial unit rather than GPS. It matters when GPS is unreliable: under a dense canopy, close to buildings or terraces, or indoors. SLAM keeps the drone positioned and helps it avoid obstacles in exactly the tight, signal-poor spots where a survey is hardest to fly.

Match the sensor to the question, not the other way around.

• Mapping, counting, damage — RGB.
• Crop health and stress — multispectral (NDVI / NDRE).
• Water stress and irrigation — thermal.
• Canopy structure, height, biomass — LiDAR.
• Flying under canopy or where GPS is weak — SLAM.
• Fine disease or nutrient discrimination — hyperspectral.

Most operations start with RGB and multispectral and add the rest as the work demands. Plenty of surveys carry more than one sensor on a single flight.

What sensors do drones use for crop health?

Mainly RGB cameras, multispectral sensors, and thermal cameras, with LiDAR for structure and SLAM for positioning where GPS is weak. Multispectral is what produces NDVI.

Is NDVI a sensor?

No. NDVI is an index calculated from a multispectral sensor's red and near-infrared bands — not a sensor in its own right.

Which sensor is best for crop health?

Multispectral, because NDVI and NDRE reveal stress before it's visible. RGB and thermal complement it rather than replace it.

What is SLAM on a drone?

Simultaneous Localization and Mapping — a technique that lets a drone position itself and map in real time without GPS, useful under canopy or near structures.

Do I need all of these sensors?

No. Most operators start with RGB and multispectral and add thermal, LiDAR, or SLAM as specific jobs call for them.

Thermal or multispectral — what's the difference?

Multispectral reads plant vigour and health; thermal reads canopy temperature and water stress. They answer different questions and work well together.

No single sensor sees everything. RGB maps the field, multispectral reads its health, thermal finds where it's thirsty, LiDAR measures its shape, and SLAM keeps the drone flying where GPS can't reach. The best surveys pick the sensor — or the combination — that answers the question in front of them.

For where these surveys fit in practice, see our guide to agricultural drones in the Philippines, plus the deep dives on NDVI mapping and LiDAR for canopy health. New to flying? Practise survey missions in our free drone simulator first.

Lumipad Drones is a non-profit that trains rural Filipinos to build, fly, and maintain low-cost agricultural drones, and to launch the microenterprises that serve local farmers. To learn more about our work, see our about page, or apply to join a program. You can also try our free drone flight simulator — built for agriculture and the Philippines, and runnable right in your browser.