Yes, but only partly: a LiDAR sensor measures depth with its own light, while the camera that tracks your phone's position and records color still needs to see detail. In a dim room you can get rough geometry with drift and a muddy texture, so adding soft, even light almost always beats fighting the dark.
How much light you need depends on how you scan, because a LiDAR scan and a photo-based (photogrammetry) scan lean on light in different places.
What needs light during a 3D scan?
A phone scan does three jobs at once, and at most one of them, measuring depth with LiDAR, can work without room light.
Measuring depth. A LiDAR sensor sends out its own light and measures how long it takes to come back. When Apple introduced the iPhone 12 Pro on 13 October 2020, it described the new LiDAR Scanner as measuring light distance and said it improves autofocus by 6x in low-light scenes. In other words, the sensor keeps working where the regular camera starts to struggle.
Tracking where the phone is. To stitch depth frames into one model, the app has to know where the phone was for every frame. Apple's ARKit documentation explains that world tracking combines the device's motion sensors with analysis of the camera image, and that tracking quality is reduced when the camera can't see details, such as when the scene is too dark. Google's ARCore has a dedicated failure state for exactly this: INSUFFICIENT_LIGHT, "motion tracking lost due to poor lighting conditions", with the advice to ask the user to move to a more brightly lit area (reference last updated on 31 October 2024).
Recording color and texture. The colors on a scanned mesh come from camera frames. A dark, grainy frame produces a dark, grainy texture, however good the depth was.
So the useful question is never only "does the sensor work in the dark". It is whether the weakest of those three jobs still works in the light you have.
Does LiDAR scanning work in the dark?
The depth measurement holds up better than the camera does, but a LiDAR scan is not a camera-free scan. Scanning apps fuse LiDAR depth with camera-based tracking, so once the room is dark enough for tracking to degrade, depth frames start landing in the wrong place. You see it as walls that double up, edges that smear, or a model that slowly bends as you walk around the object.
Color is the other casualty. A room scanned with the lights dimmed can come out with plausible proportions and an almost black texture. That may be acceptable for a rough check of a shape, but it is useless for a product page or a client presentation.
If you are not sure which sensor on your phone does what, LiDAR vs TrueDepth Camera for 3D Scanning on iPhone explains the difference between the rear LiDAR and the front face-scanning camera.
Why does photogrammetry fail in dim light?
Photogrammetry uses no depth sensor at all. It builds the model by finding the same points in many overlapping photos, which is also how you 3D scan without LiDAR on an iPhone. Everything therefore depends on image quality, and low light attacks image quality from three sides.
Noise from high ISO. When light drops, the camera raises its sensitivity. The Agisoft Metashape user manual (version 2.1, published 2024) says ISO should be set to the lowest value, because high ISO values add noise to the images. Noise is random grain that changes from frame to frame, so the software finds fewer real points that it can match between photos.
Blur from slow shutter speeds. The other way a camera compensates is by keeping the shutter open longer. The same manual warns that a shutter speed that is too slow causes blur from slight movements. Hand-held in a dim room, some blur is hard to avoid. Why 3D Scans Come Out Blurry and How to Fix It goes through the other causes of blur as well.
Missed focus. Apple's capture guide for RealityKit Object Capture notes that in low-light situations, autofocus can have difficulty finding and maintaining focus, and recommends focusing manually and putting the camera on a tripod. A photo that is soft on the object but sharp on the background hands the software the wrong points.
All three end the same way: the photos stop lining up, and you get holes, floating fragments or a scan that fails to process.
How do you scan when the room is dim?
Most low-light problems go away once you treat light as part of the setup rather than something you put up with.
- Add light before you add tricks. Apple's Object Capture guide suggests increasing ISO or adding more light when there is not enough light for a sharp photo with good depth of field. Of the two, more light is the better trade, because it fixes noise, blur and focus at once.
- Make the light soft. The same guide says hard light from an on-camera flash, direct sunlight or a bare bulb casts hard shadows that confuse the photogrammetry algorithm. It recommends bouncing light off a reflector, a wall or the ceiling, or diffusing it through a lamp shade or thin white fabric. Several ordinary lamps bounced off the ceiling usually give softer light than one bright lamp pointed straight at the object.
- Keep the light still. A light that moves with you, like the phone's own flashlight, puts shadows and highlights in a different place in every frame. Fixed lamps keep the object looking the same from shot to shot, which is what matching needs.
- Lock exposure and focus. Apple advises keeping camera settings consistent through the whole capture; on an iPhone or iPad you can long-press the camera view until "AE/AF Lock" appears. Otherwise the phone keeps changing brightness as you move between the lit and shaded sides.
- Steady the camera. For photo-based scans, a tripod and a turntable let you use a longer exposure at low ISO without blur, and a remote trigger stops the press of a button from shaking the shot. For live scanning, move slowly: ARKit's documentation says excessive motion blurs the image and reduces tracking quality, and dim light makes that blur worse.
- Take the app's warnings seriously. A message asking for more light means the tracking is losing, or has already lost, its place. Stop, fix the light and restart that pass rather than pushing through.
The same rules hold whatever app you use. MagiScan runs on iPhone and Android, and on either one the scan can only be as good as what the phone's camera and sensors see, so the time spent setting up lamps pays off.
Which scanning method copes best with low light?
| Method | What still works in dim light | What breaks first | Best low-light fix |
|---|---|---|---|
| LiDAR scan on an iPhone Pro or iPad Pro | Depth measurement, which uses the sensor's own light | Camera tracking and color texture | Soft, even room light and slow movement |
| Hand-held photogrammetry | Very little, since every step uses the photos | Focus, then blur and noise, then alignment | More light; this is the weakest option in the dark |
| Photogrammetry on a tripod or turntable | Longer exposures at low ISO, because the camera does not move | Autofocus, unless you focus manually | Manual focus, locked settings and a remote trigger |
| Camera-only AR scanning without a depth sensor | Little, since tracking and shape both come from the camera image | Motion tracking, which ARCore can stop with a low-light error | Brighter, diffused light and no blank surfaces |
In practice, a tripod-based photo capture handles a dim room best if you can control everything and have time, a LiDAR scan is the most forgiving hand-held option, and hand-held photogrammetry is the first to fall apart. ARCore's reference also lists insufficient visual features as a separate failure and asks users to avoid blank walls and surfaces without detail, which is why the last row mentions them.
Is a dark room the same problem as a dark object?
No. A dark room means too little light reaches the camera from everything in the scene. A dark object reflects little light even in a bright room, so its surface shows less detail than its surroundings do. More soft light helps with both, but the object-side problems need their own techniques. The Metashape manual tells users to avoid untextured, shiny, highly reflective or transparent objects, and to wait for a cloudy day if shiny objects cannot be avoided. How to 3D Scan Shiny Black or Transparent Objects covers those surfaces in detail.
So should you scan in low light at all?
Scan in low light only when you have no choice, and expect rough geometry rather than a finished model. When you do have a choice, spend the time on light: soft, fixed and even, with exposure and focus locked. Then move slowly, or use a tripod for photos. A clean capture is what makes the rest of the work easy, whether you export the result from MagiScan for 3D printing, AR or an e-commerce listing, or hand it to another tool.
Frequently Asked Questions
Can I use my phone's flashlight to 3D scan in the dark?
It is better than nothing, but a light that moves with the phone shifts shadows and highlights in every frame and behaves like hard light at close range. Apple's Object Capture guide warns that hard light, such as an on-camera flash, casts shadows that confuse photogrammetry. Fixed, diffused lamps give far more consistent results.
Does iPhone LiDAR scanning work in complete darkness?
The LiDAR sensor emits its own light, so its depth measurement does not depend on room light. The scan as a whole still relies on the camera for tracking and color, and Apple's ARKit documentation says tracking quality drops when the scene is too dark. Expect drift or lost tracking, and a dark or patchy texture.
What ISO should I use for photogrammetry in low light?
As low as you can. The Agisoft Metashape manual recommends the lowest ISO value because higher values add noise. Put the camera on a tripod and use a longer exposure instead, and keep the settings unchanged for the whole set of photos.
Why does my scanning app tell me to move to a brighter area?
Usually because its motion tracking is failing. Google's ARCore reports a specific INSUFFICIENT_LIGHT reason for this and suggests moving to a more brightly lit area. Fix the lighting and restart the pass rather than carrying on, because frames captured while tracking is poor tend to land in the wrong place.