NIR Camera Agriculture: How It Works and What It Shows
An NIR camera in agriculture is an imaging sensor that records near-infrared light, roughly 700 to 1,000 nm, reflected from crops and soil. Healthy leaves reflect far more of that light than visible light, and stressed plants reflect less. Compare the near-infrared image with a red or visible image, and you can map plant vigor, water stress and bare-soil conditions across a field, often before the problem shows to the eye.
The same principle works on a drone, a tractor mast or a greenhouse rail. What changes is the optics, the filters, the shutter type and how you process the data. A near infrared camera that suits a quick scouting flight is not the one you would build into a fixed monitoring station.
Vadzo Imaging designs USB, MIPI and GigE board camera modules for OEMs and system integrators, so this article looks at the topic from the hardware side. You will see why NIR matters, how the camera captures it, how to run a field workflow, and what to check when you choose a camera for a farming or drone product.
Why NIR Camera matters for agriculture crop and soil monitoring
NIR imaging matters because it shows plant health and water status that an ordinary color camera cannot, and it shows them earlier. A leaf can look green for days after its internal condition has started to change.
Why healthy plants stand out in near-infrared
Chlorophyll absorbs red and blue light for photosynthesis. The internal cell structure of a healthy leaf scatters near-infrared light back toward the sensor. A healthy canopy therefore reflects a lot of NIR and very little red.
Drought, disease and nutrient shortage change that balance. NIR reflectance falls and red reflectance rises. The standard index, NDVI, captures this as (NIR − Red) / (NIR + Red). Dense healthy vegetation typically scores about 0.6 to 0.9, while bare soil sits near 0.1 to 0.2, though exact values vary by crop and growth stage.
NIR shows how a plant is performing, not just how green it looks.
What you can monitor with it
Near-infrared camera agriculture work usually falls into a few jobs:
Crop vigor and emergence: spot thin stands, skips and weak zones after planting.
Water stress and irrigation zoning: find the parts of a field that dry out first.
Soil variation: water absorbs NIR, so wet bare soil looks darker than dry soil. That helps you map moisture patterns before planting.
Produce quality: indoor systems use NIR to check freshness and bruising on sorting lines and in vertical farms.
How an NIR camera captures near-infrared light
Standard silicon CMOS sensors already respond to light up to about 1,000 nm. A color camera blocks that range with an IR-cut filter. An NIR camera removes or replaces that filter, and usually adds a bandpass filter so only the band you want reaches the sensor.
Sensor, filter and illumination
A monochrome sensor has no color mosaic, so it collects more light per pixel and gives a cleaner NIR image. A bandpass filter, often centered near 850 nm, isolates the band you want to measure. Outdoors, sunlight supplies plenty of NIR. Indoors or at night, you add 850 nm LED lighting.
Past about 1,000 nm, silicon sensitivity fades. Shortwave infrared needs different sensor materials and costs far more, so most farm systems stay in the 700 to 1,000 nm range.
If your question can be answered below 1,000 nm, a silicon sensor and a filter will do the job.
Camera setups compared
Setup | What it captures | Best for | Main tradeoff |
|---|---|---|---|
Modified color camera | Visible and NIR mixed, with a filter swap | Low-cost scouting | Overlapping bands, lower accuracy |
Mono camera plus NIR bandpass | One clean NIR band | Fixed stations, custom payloads | Needs a second camera for red |
RGB-NIR sensor | Red, green, blue and NIR from one sensor | Compact single-camera designs | Lower resolution per band |
Multispectral array | Several narrow bands on separate sensors | Research-grade vegetation indices | Cost, weight, synchronization |
Hyperspectral | Dozens to hundreds of narrow bands | Produce freshness and research | Cost, data volume |
How to use an NIR camera for crop and soil monitoring
You use an NIR camera by fixing the question first, then capturing calibrated images under consistent light and checking the result on the ground. The sequence matters more than the hardware.
A field workflow in six steps
Define the question. Emergence, irrigation zones and nutrient problems need different timing.
Choose the platform. Use a handheld or mast mount for small plots and a drone for whole fields. In the US, commercial drone flights generally fall under FAA Part 107.
Pick the bands. You need NIR plus red at minimum to calculate NDVI.
Calibrate. Photograph a reflectance panel before each flight, and fly near solar noon under similar sky conditions each time.
Capture with overlap. About 75% front and 70% side overlap gives stitching software enough to work with.
Process and verify. Stitch an orthomosaic, calculate the index, then walk the low-scoring zones before you act.
Worked example: ground resolution from a drone
Ground sample distance equals pixel size times altitude, divided by focal length. Take a 20MP mono sensor (5,472 × 3,648 pixels) with 2.4 µm pixels and an 8 mm lens, flown at 100 m. That gives 0.0000024 × 100 / 0.008 = 0.03 m, so 3 cm per pixel. Each image covers about 164 m by 109 m.
Drop to 50 m and you get 1.5 cm per pixel, but you fly more passes and capture more images. For field-level zoning, several centimeters per pixel is usually enough. Row-level work needs finer detail.
Limits that trip people up
Clouds change illumination mid-flight, so a calibration panel is not optional. NDVI also saturates in dense canopy, which hides differences between good and excellent stands. And NIR flags a symptom, not a cause. A low value tells you where to look, not whether the culprit is water, disease or nitrogen.
Treat every NIR map as a scouting guide, and confirm on the ground before you spray, irrigate or fertilize.
Choosing NIR camera systems for farming and drones
Choose by platform first, then by shutter, NIR sensitivity, interface, weight and environmental range. Cost follows from those choices. Bare board modules start in the low hundreds of dollars, while packaged multispectral systems cost considerably more because they add calibrated optics, several sensors and processing software.
Spec checklist
Spec | Why it matters in the field |
|---|---|
Global shutter | Avoids image skew at drone speeds. Rolling shutter suits fixed mounts or slow flights |
NIR sensitivity at 850 nm | Sets signal strength in your band, so check the quantum efficiency curve |
Filter mounting | You need a removable IR-cut filter or a place for a bandpass filter |
Interface | MIPI for onboard Jetson or Raspberry Pi processing, USB UVC for plug-and-play, GigE with PoE for cable runs up to 100 m |
Weight and size | Drone payload and battery life depend on it |
Operating temperature | Outdoor units face heat, cold and vibration |
Where Vadzo board cameras fit
Vadzo does not sell a packaged agricultural multispectral camera. It supplies board-level modules that an OEM builds into its own payload. Its custom engineering covers lens and filter integration, electro-mechanical lens filter control and NIR LED array boards, which are the pieces an NIR product needs.
Model | Interface | Sensor | Shutter | Notes |
|---|---|---|---|---|
Falcon-2020MRS | USB 3.2 Gen 1 | onsemi AR2020, 20MP mono | Rolling | Enhanced NIR response at 850 and 900 nm |
Bolt-1335MRS | MIPI CSI-2 | onsemi AR1335, 13MP mono | Rolling | 13 g without lens, from $295 |
Innova-234MGS | GigE with PoE | onsemi AR0234, 2.3MP mono | Global | 25 g without lens, from $418 |
Thermal imaging is a different band. The Quanta-234CGE pairs a 2.3MP global shutter visible sensor with a 256 × 192 LWIR thermal module at 8 to 14 µm, for temperature measurement rather than NIR reflectance. At $900, it suits canopy temperature or equipment monitoring alongside an NIR camera.
Match the shutter and interface to your platform first, and the sensor to your band second.
Start with the question, then pick the camera
An NIR camera turns reflected near-infrared light into a map of plant vigor, water stress and soil variation. Silicon sensors, a filter and a calibrated workflow are enough for most farm questions. The hard part is consistency: the same light, the same altitude, the same calibration, and a walk through the field to confirm what the map suggests.
When you choose hardware, decide the platform and the band first. Then check shutter type, NIR sensitivity, interface and weight against that platform. If you are building your own drone payload or monitoring station and want a camera tuned for it, explore Vadzo's board camera modules and custom engineering to find a starting point that fits your design.
