What a 3D Scanner Does

A 3D scanner is a device that measures the shape and surface of a physical object and converts that information into a digital 3D model you can view on a computer, edit, print, or manufacture. Instead of taking a flat photograph, it records depth—how far away each point on the object's surface is from the scanner. The result is a complete digital copy of the object's geometry.

The scanner does not need to touch the object. It sends out light or sound waves, measures what bounces back, and uses that data to build a point cloud—a collection of millions of tiny dots that define the object's shape. Software then connects these dots into a smooth surface model that designers, engineers, and manufacturers can work with.

Key Takeaways

  • 3D scanners work by sending out light or sound, measuring the reflections, and recording the distance to each point on an object's surface.
  • The scanner creates a point cloud—millions of coordinate points in 3D space—which software then converts into a usable 3D model.
  • Different scanner types (laser, structured light, photogrammetry, and time-of-flight) work best for different object sizes, materials, and accuracy needs.
  • Shiny, transparent, and very dark surfaces can confuse scanners because they reflect or absorb light unpredictably.
  • After scanning, the raw data usually needs cleaning and processing before it becomes a finished model ready for design or manufacturing.

How Light-Based Scanners Measure Distance

Most 3D scanners use light to measure distance. A laser scanner shoots a thin laser beam at the object and measures how long the light takes to bounce back. Since light travels at a known speed, the scanner can calculate the exact distance. It repeats this thousands or millions of times, moving the laser across the entire object's surface to build a complete map.

Structured light scanners work differently. Instead of a single laser beam, they project a pattern of light—often a grid or series of stripes—onto the object. A camera watches how the pattern bends and stretches across the surface. Where the object curves or has a bump, the pattern distorts. The scanner measures that distortion to figure out the depth at each point. This method is faster than laser scanning because it captures many points at once instead of one at a time.

Both methods create a point cloud: a list of millions of coordinates (X, Y, Z positions) that describe where the surface is in 3D space. The denser the point cloud, the more detail the final model will have.

Photogrammetry: Using Photographs Instead of Lasers

Photogrammetry is a different approach that does not use a dedicated scanner device. Instead, you take many overlapping photographs of an object from different angles—often 50 to 200 photos depending on the object's complexity. Software analyzes the photos, finds matching features in overlapping images, and calculates the 3D position of each feature based on how it appears from multiple viewpoints.

This method works well for large objects, outdoor scenes, and situations where you cannot bring a scanner to the location. A smartphone camera can do basic photogrammetry with the right software. The trade-off is that photogrammetry is slower than laser or structured light scanning, and it works best on objects with visible texture and detail—a blank white sphere is nearly impossible to scan this way.

Time-of-Flight Scanners and Depth Cameras

Time-of-flight scanners measure how long it takes light to travel to an object and back, similar to laser scanners but using infrared light instead of visible lasers. They can scan an entire scene at once rather than point by point, making them fast enough for real-time applications like motion capture or augmented reality filters on smartphones.

Depth cameras—the kind built into some tablets and gaming consoles—use time-of-flight or a related method called structured infrared light. They are not as precise as dedicated laser scanners, but they are compact, affordable, and fast enough for many uses. They work well for scanning people, faces, and rooms, but struggle with shiny surfaces and outdoor sunlight.

Why Some Materials Are Hard to Scan

Not all surfaces behave the same way with light. Shiny or reflective surfaces—like polished metal, glass, or mirrors—bounce light in unpredictable directions. The scanner expects light to bounce straight back, but reflective surfaces scatter it, so the scanner either gets no return signal or a confusing one. Transparent materials like clear plastic or glass are even worse because light passes through them instead of bouncing back.

Very dark or black surfaces absorb most light instead of reflecting it, so the scanner receives a weak signal. Textured surfaces with fine detail can also cause problems because the scanner may not have enough resolution to capture the texture accurately, or the texture creates shadows that confuse the depth measurement.

To handle these difficult materials, operators sometimes spray the object with a temporary white powder or matte coating that reflects light predictably. This coating washes off and does not damage the object.

From Raw Data to Finished Model

When a scanner finishes, the raw output is a point cloud—millions of unorganized 3D coordinates. This data is messy: it may have gaps where the scanner could not reach, noise from reflections or vibration, and outlier points that are clearly wrong. Software must clean this up before the model is useful.

The first step is registration: if you scanned the object from multiple angles, the software aligns all the point clouds so they overlap correctly. Next comes noise reduction, which removes stray points and smooths out measurement errors. Then the software creates a mesh—a surface made of triangles or polygons that connects the point cloud into a solid shape.

Finally, the mesh may be refined, smoothed, or edited by hand to remove artifacts or fill small holes. The result is a 3D model file (usually in formats like STL, OBJ, or PLY) that can be imported into design software, 3D printers, or game engines.

Choosing the Right Scanner Type for Your Object

The best scanner depends on what you are scanning. For small manufactured parts that need high precision, a laser scanner is the standard—it can achieve accuracy within fractions of a millimeter. For larger objects or when speed matters, structured light is often faster and still accurate. For outdoor scenes, buildings, or situations where portability is critical, photogrammetry with a camera is practical and inexpensive.

For real-time applications like motion capture or face scanning, time-of-flight or depth cameras are the only choice because they work at video frame rates. If you are scanning people or organic shapes, depth cameras built into tablets or phones can work for casual projects, though professional motion capture systems are far more accurate.

Material matters too. Shiny or transparent objects are hardest to scan with any method; if you must scan them, laser scanners often perform better than structured light, but you may still need to coat the surface. Matte, textured surfaces are easiest for any scanner type.

Frequently Asked Questions

Can I scan something with my smartphone?

Yes, if your phone has a depth camera or LiDAR sensor (common on newer iPhones and some Android phones). Apps like Polycam or Scaniverse use these sensors to capture 3D models. The accuracy is lower than dedicated scanners, but good enough for casual projects, room layouts, or objects a few feet across. Older phones without depth sensors cannot do true 3D scanning, though photogrammetry apps can work if you take many photos.

How long does it take to scan an object?

A small object with a structured light scanner might take 30 seconds to a few minutes. A laser scanner on the same object could take 10 to 30 minutes because it scans point by point. Photogrammetry depends on how many photos you take and how powerful your computer is—anywhere from minutes to hours. Time-of-flight and depth cameras are the fastest, capturing in real time.

What file format should I use for a 3D scan?

STL (Stereolithography) is the most common format for 3D printing. OBJ (Wavefront) is widely supported by design software and game engines. PLY (Polygon File Format) preserves color and texture information. For professional work, formats like E57 or XYZ store the raw point cloud. Ask what format your intended software needs before you scan.

Can I scan a moving person?

Yes, but only with time-of-flight or depth cameras that capture at video speed. Laser and structured light scanners are too slow—the person will blur or move between scans. Professional motion capture systems use multiple depth cameras or infrared markers to track movement in real time. For a still portrait, any scanner type works, but the person must hold completely still for 30 seconds to a few minutes.

Why is my scan full of holes and gaps?

Holes appear where the scanner could not reach—inside crevices, under overhangs, or on the back side of the object. You must scan from multiple angles and register the scans together to fill these gaps. Some software can fill small holes automatically, but large missing areas require additional scans from different positions. Reflective or dark surfaces also create gaps because the scanner receives no signal from those areas.