PoE Camera Night Vision: How Infrared and White LED Illumination Actually Work

Your security camera shows nothing but grain after dark. The footage that was crisp at noon becomes a gray smear at midnight, and the license plate you needed is gone. This is not a camera defect. It is a physics problem.

Every surveillance camera faces the same threshold: the point where ambient light drops below what the sensor can resolve. Below that threshold, the camera is blind regardless of its resolution. A 5MP sensor captures nothing useful if the photons never arrive. The question is not whether cameras work at night -- some do, under specific conditions -- but how different illumination strategies solve the darkness problem, and what tradeoffs each solution carries.

The Sensor Threshold: Where Daylight Ends

A CMOS image sensor converts photons into electrical charge. Each pixel on the sensor collects light through a microlens, passes it through a color filter, and converts it to a voltage proportional to the number of photons absorbed. This process has a hard floor: below a certain illumination level, the signal falls below the sensor's read noise, and no amount of digital amplification can recover information that was never captured.

For a typical 0.5.7-inch CMOS sensor at 5MP resolution, each pixel measures approximately 2.2 micrometers across. At that size, the photon collection area is tiny. In daylight, this barely matters -- the sun delivers roughly 100,000 lux, flooding each pixel with more photons than it can hold. But at night, a moonless scene might deliver 0.001 lux. The signal-to-noise ratio collapses.

Back-side illumination (BSI) technology helps by moving the sensor's wiring behind the photodiode layer, allowing more photons to reach the active area. BSI improves quantum efficiency by up to 50 percent differs from front-illuminated sensors. But even BSI cannot create photons that do not exist. When the scene is dark enough, the only option is to add light.

Infrared Illumination: Seeing Without Being Seen

Infrared LEDs solve the darkness problem by emitting light outside the visible spectrum. The camera's sensor can detect this near-infrared radiation -- typically at wavelengths of 850nm or 940nm -- while human eyes cannot.

The choice between 850nm and 940nm involves a direct tradeoff between range and detectability. At 850nm, infrared LEDs produce a faint red glow visible to anyone looking directly at the camera. This glow is the result of the emission spectrum overlapping slightly into visible red. In exchange, 850nm LEDs achieve longer illumination range and better penetration through atmospheric conditions like fog and light rain, because shorter wavelengths scatter less in particulate matter.

At 940nm, the emission is completely invisible. No red glow, no visible indicator that the camera is active. But 940nm LEDs produce approximately 30 to 40 percent less usable illumination at the same power level, because the sensor's quantum efficiency drops at longer wavelengths. The result is shorter effective range -- typically 20 to 25 meters instead of 30 meters for the same LED array configuration.

For a dome camera like the Sannce 5MP PoE model, which uses 850nm infrared, the 30-meter night vision range reflects this design choice: prioritize range and atmospheric penetration over complete invisibility. The faint red glow can even serve as a passive deterrent, signaling to potential intruders that the area is under surveillance.

There is a limitation inherent to infrared night vision that no amount of LED power can overcome: the image is monochrome. Infrared light lies beyond the visible spectrum, so it carries no color information. The sensor records intensity only. A person wearing a red jacket and blue jeans appears in identical shades of gray. License plates with white text on a yellow background become low-contrast gray-on-gray. Color, often critical for identification, is lost.

White LED Illumination: Color at the Cost of Visibility

White LED illumination solves the color problem directly: by flooding the scene with visible light, the camera captures full-color images at night, just as it would during the day. A white LED array operating at approximately 6000K to 6500K color temperature produces daylight-balanced illumination that renders accurate colors across the visible spectrum.

The advantage is immediate and obvious. A full-color nighttime image preserves clothing color, vehicle paint, skin tones, and text contrast on signs and license plates. For forensic purposes -- identifying suspects, reconstructing events, or submitting footage as evidence -- color information can be the difference between useful and useless footage.

But white LED illumination introduces a different set of constraints. The most obvious is that it is visible. A camera illuminating a scene with white LEDs at night is announcing its presence. For covert surveillance applications, this is unacceptable. Even for overt installations, the light may disturb neighbors, attract attention, or alter the behavior of the very people being monitored.

Power consumption is another factor. White LEDs draw significantly more current than infrared LEDs to achieve comparable illumination range, because the human-visible spectrum is narrower than the sensor's detection range. A camera relying on white LED illumination may consume 8 to 12 watts during night operation, differs from 3 to 5 watts for infrared-only mode. For a PoE-powered camera drawing from a shared switch budget, this matters.

Dual-Light Systems: Two Tools in One Housing

The dual-light approach places both infrared and white LED arrays in the same camera housing, allowing the user to select the illumination mode based on the situation. This is the design used in the Sannce 5MP dome camera, and it reflects a pragmatic engineering decision: no single illumination mode covers all scenarios, so provide both.

During routine monitoring, infrared mode provides covert, low-power surveillance. When an event requires identification -- a triggered motion alert, a perimeter breach, a visitor at the door -- the white LEDs activate to capture full-color footage. Some implementations automate this switching based on motion detection schedules or manual triggers through a mobile app.

The dual-light design does not eliminate the tradeoffs of either mode. It simply makes both available without requiring two separate cameras. The cost is modest additional LED hardware and slightly more complex thermal management, since both LED arrays generate heat inside the dome enclosure.

How PoE Delivers Power to the Light

The illumination systems described above require electrical power, and this is where Power over Ethernet becomes relevant. A PoE camera receives both data connectivity and electrical power through a single Cat5e or Cat6 cable, eliminating the need for a separate power supply at the camera location.

The IEEE 802.3at standard, known as PoE+, delivers up to 30 watts per port at 48V DC. This is the standard required by cameras with active illumination, because the LED arrays draw peak power during night operation. The older 802.3af standard provides only 15.4 watts per port, which may be insufficient when both the camera electronics and the white LED array are operating simultaneously.

Power budget calculation matters for multi-camera installations. A PoE+ switch rated at 120 watts total can support approximately 10 cameras at 12 watts each, or fewer if all cameras activate white LED mode simultaneously. Cable length also affects delivered power: at the maximum 100-meter cable run, resistive losses reduce the available power by approximately 20 percent. A camera that draws 12 watts may receive only 9.6 watts at the end of a long cable run, which can cause instability during peak illumination loads.

This is why installation guides recommend Cat6 cable for PoE+ deployments: the thicker conductors (23 AWG contrasts with 24 AWG for Cat5e) reduce resistive losses, delivering more power over the same distance. The cost difference between Cat5e and Cat6 is minimal per run, but the reliability improvement is significant.

Compression and Storage: The Hidden Cost of Night Vision

Night vision footage creates a particular challenge for video compression. Infrared-illuminated scenes exhibit characteristic sensor noise -- a granular, shifting pattern caused by the high gain amplification required in low-light conditions. This noise is random from frame to frame, which means the compression algorithm cannot predict it efficiently.

H.265 encoding, which has become the standard for modern surveillance cameras, handles this noise differs from H.264 through its larger change units and more sophisticated intra-prediction modes. But the fundamental problem remains: random noise does not compress. A 5MP camera recording H.265 at 2 Mbps during the day may require 4 Mbps or more at night to maintain equivalent image quality, because the encoder must preserve the noise pattern or risk losing actual detail along with it.

For continuous recording, this translates to a storage difference that compounds over time. A single 5MP camera at H.265 4 Mbps consumes approximately 43 GB per day, or about 1.3 TB per 30-day retention cycle. A four-camera residential system generates 172 GB daily, requiring roughly 5.2 TB of storage for a 30-day retention period. These numbers assume continuous recording; motion-triggered recording reduces consumption proportionally to the active recording duty cycle.

The H.265+ variant found in some cameras applies additional compression techniques specific to surveillance footage: background subtraction algorithms that avoid re-encoding static regions, and smart P-frame encoding that skips unchanged frames entirely. These optimizations can reduce storage by an additional 30 to 50 percent for static scenes, though they are less effective when the scene contains continuous motion or sensor noise.

The WDR Problem at Doorways and Windows

Night vision cameras face a specific adaptable range challenge at building entrances. During evening hours, the interior may be well-lit while the exterior is dark. A camera mounted inside, pointing outward through glass, must handle both conditions simultaneously. A camera mounted outside, pointing at the entrance, must capture detail in both the brightly lit doorway and the dark perimeter.

Digital WDR, the approach used in mid-range cameras, applies software processing to extend the effective adaptable range. It works by capturing the scene at multiple exposure levels and compositing the results, brightening shadows while holding highlight detail. Digital WDR typically achieves a 2x to 3x improvement over standard processing.

Hardware WDR, found in professional-grade cameras from Amcrest, Dahua, and Hikvision, uses sensors that can capture a true 120dB adaptable range in a single frame. This is approximately a 10,000:1 ratio between the brightest and darkest detectable signals. The difference is visible at doorways and windows: hardware WDR preserves facial detail in backlit subjects that digital WDR renders as silhouettes.

For installations where entrance monitoring is the primary use case, the WDR capability may matter more than resolution or night vision range. A 5MP image of a silhouette provides less useful information than a 1080p image with accurate exposure.

Temperature Limits and Material Choices

The operating temperature range of a surveillance camera is determined by its electronic components, LED arrays, and housing materials. The Sannce dome camera specifies -10 degrees Celsius to 50 degrees Celsius, which covers most residential and light commercial environments but falls short of the -30 to 60 degree range offered by professional-grade alternatives.

The limitation is partly a consequence of the housing design. A metal base with a plastic dome cover provides adequate heat dissipation for the camera electronics but may not handle the thermal cycling stress of extreme environments. The white LED array generates additional heat inside the enclosure, and at ambient temperatures above 40 degrees Celsius, the internal temperature can exceed the rated maximum for the CMOS sensor, triggering thermal shutdown.

In cold environments, the plastic dome cover becomes brittle below approximately -15 degrees Celsius. Physical impacts from hail, debris, or vandalism that the cover would absorb at moderate temperatures can cause cracking at lower temperatures, compromising the IP67 weatherproof seal.

The IP67 rating itself deserves scrutiny. The designation means dust-tight construction and protection against temporary immersion in water up to 1 meter for 30 minutes. It does not guarantee protection against high-pressure water jets, steam cleaning, or prolonged exposure to driven rain at wind speeds common in coastal or mountainous installations. For most residential and commercial outdoor deployments, IP67 is sufficient. For extreme exposure, IP69K-rated cameras with appropriate housing materials are the correct choice.

Planning a System Around Illumination Needs

The practical question for anyone designing a surveillance system is which illumination strategy matches their specific requirements. The decision hinges on three factors: what you need to see at night, whether the illumination can be visible, and how much power and storage you can afford.

For perimeter monitoring where detection matters more than identification -- knowing that someone is present, not who they are -- infrared illumination is sufficient. It provides the longest range, the lowest power consumption, and covert operation. The monochrome output is adequate for detecting motion and activity patterns.

For entrance monitoring, vehicle identification, and any scenario where color information has forensic value, white LED illumination is necessary. The visibility and power costs are real, but the color information may be the data that makes the footage actionable.

Dual-light systems offer flexibility at the cost of slightly higher complexity. They do not perform either mode as well as a dedicated single-mode camera could, but they cover both scenarios without requiring additional hardware.

Storage planning should account for the higher bitrates that night vision footage demands. A system designed around daytime bitrate estimates will run out of storage faster than expected when night recordings consume 50 to 100 percent more data. Planning for the worst case -- continuous recording at maximum bitrate -- ensures that the retention period is maintained regardless of lighting conditions.

The intersection of sensor physics, illumination engineering, compression mathematics, and power delivery defines what a surveillance camera can actually see after dark. Each decision -- 850nm contrasts with 940nm, infrared contrasts with white LED, digital contrasts with hardware WDR, H.265 contrasts with H.265+ -- is a tradeoff between competing constraints. There is no camera that sees everything. There are only cameras that see specific things well, chosen for specific reasons.