What a 3D scanner does and how it differs from a camera

A 3D scanner captures the shape and surface of a physical object and converts that information into a digital three-dimensional model. Unlike a camera, which records only what it sees from one angle as a flat image, a 3D scanner measures depth — the distance from the scanner to every point on the object's surface. The result is a file that contains coordinates for thousands or millions of points in space, which software can then render as a solid shape you can rotate, measure, and modify on screen.

The scanner does not care about color or texture the way a camera does. Its job is geometry: where is the surface, and how far away is it? Some 3D scanners do capture color information alongside the depth data, but that is secondary. What matters is the three-dimensional map.

Key Takeaways

  • 3D scanners measure depth by sending light or sound waves at an object and timing how long the signal takes to return, or by analyzing how the object distorts a known pattern of light.
  • Structured light scanners project a grid or stripe pattern onto an object and use cameras to see how the pattern bends, which reveals the object's shape.
  • Laser scanners bounce a laser beam off the surface and measure the distance based on how long the light takes to return, making them precise but slower than structured light.
  • Photogrammetry uses overlapping photographs from different angles and software that finds matching details between images to calculate depth and build a 3D model.
  • The output is a point cloud or mesh file that contains thousands of coordinate points; software then smooths and refines this data into a usable 3D model.

Structured light: projecting a pattern and reading the distortion

The most common type of 3D scanner in consumer and industrial settings is a structured light scanner. It works by projecting a known pattern — usually a grid of lines or a series of stripes — onto the object. A camera (or multiple cameras) watches how that pattern bends and warps as it hits the object's surface. Where the object is closer to the scanner, the pattern compresses; where it is farther away, the pattern stretches. The scanner's software analyzes these distortions and calculates the depth at each point.

This method is fast and works well on objects with varied surfaces — matte finishes, slight curves, textured materials. It struggles with shiny or reflective surfaces because the light bounces unpredictably instead of following the pattern. It also has trouble with transparent or translucent materials like glass or clear plastic, because light passes through them rather than reflecting back.

Structured light scanners are common in 3D printing workflows, quality control in manufacturing, and dental scanning. They typically scan an object in seconds to a few minutes, depending on the size and complexity.

Laser scanners: timing the return of a bounced beam

A laser scanner works on a different principle: it fires a laser beam at the object and measures how long the light takes to bounce back. This is called time-of-flight measurement. The scanner knows the speed of light, so if the beam takes 10 nanoseconds to return, the scanner can calculate that the surface is roughly 1.5 meters away. By sweeping the laser across the object or rotating the scanner, it builds up a point cloud of distance measurements.

Laser scanners are very accurate and work on almost any material, including shiny or dark surfaces. They also work at longer distances than structured light scanners — some can scan objects 100 meters away. The trade-off is speed: laser scanners are slower because they measure one point at a time (or a line at a time), whereas structured light scanners capture an entire pattern in one shot.

Laser scanners are common in surveying, architecture, and large-scale industrial inspection. They are also used in autonomous vehicles and robotics to map surroundings in real time.

Photogrammetry: building 3D models from overlapping photographs

Photogrammetry is a different approach altogether. Instead of using a dedicated scanner, it uses a regular camera or smartphone to take many photographs of an object from different angles. Software then analyzes these images to find matching details — a corner, a texture, a shadow — that appear in multiple photos. By calculating where those details must be in three-dimensional space to appear in all those different images, the software reconstructs the object's shape.

Photogrammetry does not require special hardware, which makes it cheap to start. It works well on textured objects with visible details — a sculpture, a building, a face. It struggles with featureless surfaces like a blank white wall or a shiny sphere, because the software has nothing to match between images.

The process is slower than structured light scanning because the software has to analyze dozens or hundreds of images. But the results can be very detailed, and photogrammetry captures color and texture naturally as part of the photograph data.

What happens to the data after scanning

The raw output of any 3D scanner is a point cloud — a file containing millions of coordinate points (X, Y, Z positions) that represent the surface of the object. This point cloud is noisy and unorganized; it is the digital equivalent of a rough sketch. Software must then process this data to create a usable model.

The first step is usually mesh generation: the software connects the points with triangles to create a solid surface. This mesh can then be smoothed, cleaned up, and refined. Holes are filled, stray points are removed, and the surface is simplified to reduce file size. The result is a 3D model file — typically in formats like STL, OBJ, or PLY — that can be imported into design software, 3D printing software, or game engines.

Some scanners include software that does much of this processing automatically. Others output raw point clouds and leave the refinement to the user. The quality of the final model depends on both the scanner's accuracy and the skill of the person processing the data.

Accuracy, resolution, and what they mean for your use

Accuracy is how close the scanner's measurements are to the true dimensions of the object. A scanner with 0.1 mm accuracy will measure distances within one-tenth of a millimeter. Resolution is how many points the scanner captures per unit area — higher resolution means more detail and a finer mesh.

For 3D printing a decorative object, accuracy of 1 to 5 mm is usually fine. For manufacturing a part that must fit with other parts, you might need 0.1 mm or better. Photogrammetry can achieve high resolution but is often less accurate than laser or structured light scanners because it depends on how well the software can match details between images.

Scan time, file size, and processing time all increase with resolution. A quick scan for rough visualization might take 30 seconds; a high-resolution scan for manufacturing might take 10 minutes and produce a file gigabytes in size.

Common uses for 3D scanning

In manufacturing and quality control, 3D scanners compare finished parts to a reference model to catch defects. In dentistry and orthopedics, scanners capture the shape of teeth or bone to design custom implants or braces. In archaeology and museum work, scanners digitize artifacts so they can be studied, archived, and even 3D-printed for handling without damaging the original.

In entertainment and game development, photogrammetry is used to scan real actors, props, and environments and convert them into digital assets. In construction and surveying, laser scanners map buildings and terrain to create as-built documentation and detect changes over time.

For hobbyists, 3D scanners are used to digitize small objects for 3D printing, to create models for video games or animation, or straightforward to experiment with the technology. The barrier to entry has dropped significantly as smartphone-based photogrammetry and affordable structured light scanners have become available.

Frequently Asked Questions

Can a 3D scanner work on shiny or reflective objects?

Structured light scanners struggle with reflective surfaces because the light bounces away instead of following the projected pattern. Laser scanners handle reflective materials better because they measure the time it takes light to return, which works even if the reflection is imperfect. Photogrammetry also works on shiny objects if they have visible texture or detail. Many users spray reflective objects with a temporary matte coating before scanning.

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

STL (stereolithography) is the most common format for 3D printing. OBJ is widely supported by design and animation software and includes color and texture data. PLY (polygon file format) also stores color and is good for detailed scans. The choice depends on what software you plan to use next. Most 3D scanners can export to multiple formats.

How long does it take to scan an object?

Structured light scanners typically take 10 seconds to a few minutes. Laser scanners take longer — anywhere from a few minutes to an hour for large or complex objects. Photogrammetry depends on how many photos you take and how powerful your computer is; processing can take 30 minutes to several hours. The final model usually needs additional cleanup and refinement regardless of the method.

Do I need special lighting for 3D scanning?

Structured light and laser scanners work in various lighting conditions because they produce their own light source. Photogrammetry is more sensitive to lighting because it relies on photographs; even, diffuse lighting without harsh shadows usually produces better results than direct sunlight or spotlights.

Can a smartphone scan 3D objects?

Yes, using photogrammetry apps or LiDAR sensors built into newer phones. The results are less precise than dedicated scanners but good enough for casual use, 3D printing, or creating game assets. Apps like Polycam and Scaniverse use the phone's camera and motion sensors to build 3D models as you walk around an object.