Hardware · Sensor family

Cameras & imaging.

What these are
Eyes on the crop: photos, video, and vision
Uses
Time-lapse, check-ins, pests, AI, security
Open Agriculture Technology take
Match the camera to the job, and keep it local

A camera is a sensor that reads what a number cannot: a wilting leaf, a pest on the underside, a door left open, the slow march of a time-lapse. For a grower it is the cheapest way to be in two places at once. The trick is matching the camera to the job, because a two-dollar chip and a security camera solve very different problems.

A Raspberry Pi camera module
Image: adafruit.com

A camera is a sensor.

The other sensors here give you a single number; a camera gives you the whole scene, and increasingly a number on top of it. A time-lapse turns a season into a thirty-second clip. A remote view answers “is the greenhouse all right” from the kitchen. And paired with a little computing, a camera can count: spotting pests, flagging disease, or measuring canopy cover. That last step, plant vision, is where a camera stops being a convenience and becomes a real instrument.

Compare the cameras.

Four ways to put an eye on the crop, from a few dollars to a polished appliance. The prose under the table is the guide, since each suits a different job.

Cameras for growing · verified 2026-06-23
Spec ESP32-CAM Pi Camera USB webcam IP camera
Needs Just the board + Wi-Fi A Raspberry Pi A Pi or PC Your network
Quality Low (2 MP), basic High, autofocus Decent High
Wireless Wi-Fi built in Via the Pi Via the host Wi-Fi or PoE
Effort Fiddly to flash Moderate Easy Plug and play
Cost A few dollars Pi + ~$25 Cheap Tens to hundreds
Best for A cheap networked snapshot Quality vision and AI Simple monitoring, time-lapse Security, continuous watch

The ESP32-CAM is the cheap, all-in-one option, but it is fiddly: no native USB (you flash it through an adapter), only 2 megapixels, and a tendency to brown out on weak power. A newer ESP32-S3 camera board is friendlier. For real image quality or any AI work, a Raspberry Pi with a camera module or a plain USB webcam is the better path; for hands-off security, an IP camera.

Keep the watching local.

A camera is the most personal sensor on the farm, so where its footage goes matters. Many consumer cameras push everything to a vendor cloud by default. The own-your-data path keeps it on your own hardware: a Raspberry Pi running Frigate or Home Assistant can record local IP-camera streams and even run on-device object detection, with nothing leaving your network unless you choose. Look for cameras that speak the open RTSP stream protocol rather than cloud-only ones, and the footage stays yours.

Other watchers.

The rest of the “plant and space” corner watches without a lens. A PIR motion sensor notices movement (a person, an animal, a door). A time-of-flight or ultrasonic distance sensor measures a gap, useful for tank and reservoir levels or bench presence. A load cell weighs, which can track a plant’s water use or a hive’s honey. None of these need a camera, and all read with any microcontroller.

Where they fit, and where they don’t.

Where they fit

  • Time-lapse and remote check-ins on a crop.
  • Spotting pests, disease, or canopy changes by eye or AI.
  • Local security and continuous monitoring (IP camera).
  • Level and presence, with distance or motion sensors.

Where they don’t

  • Sending a far-field image over LoRa; it is far too much data.
  • Quality vision from an ESP32-CAM; it is low-res and fiddly.
  • Cloud-only cameras, if you care about owning the footage.
  • Battery-only AI vision; it needs steady power and a computer.

Resources.

These open in a new tab:

Raspberry Pi Camera Module 3 Frigate (local camera AI) Espressif camera (ESP32-CAM) The Raspberry Pi

Frequently asked questions.

What camera should I use for a greenhouse?

It depends on the job. For a cheap networked snapshot, an ESP32-CAM works but is fiddly and low-res. For good image quality or AI, use a Raspberry Pi with a camera module or a USB webcam. For hands-off security and continuous recording, an IP camera fed into local software.

Is the ESP32-CAM any good?

It is cheap and all-in-one, with Wi-Fi built in, which is its appeal. But it is fiddly: you flash it through a separate adapter, it is only 2 megapixels, and it can brown out on weak power. A newer ESP32-S3 camera board is easier, and a Raspberry Pi camera is far better for quality.

Can I keep my camera footage off the cloud?

Yes. Choose cameras that provide an open RTSP stream rather than cloud-only ones, and record them on your own hardware with software like Frigate or Home Assistant on a Raspberry Pi. The footage, and any AI detection, then stays on your network.

Can a camera detect pests or disease automatically?

With a computer behind it, increasingly yes. A Raspberry Pi or an ESP32-S3 can run object-detection or classification models to flag pests, disease, or canopy changes. It takes setup and good lighting, but plant vision is a real and growing use.

Can I send camera images over LoRa?

No. LoRa carries tiny messages, and an image is far too much data. For a remote camera you need Wi-Fi, cellular, or a wired link with real bandwidth. LoRa is for numbers, not pictures.