Devices that scan 2D images of 3D objects

A 3D scanner is a device that captures a flat photograph or video of a physical object and converts it into digital 3D data — a computer model you can rotate, measure, or print. The scanner does not create the 3D model by itself. Instead, it takes 2D images (ordinary photographs) from different angles, then uses software to figure out the object's shape, depth, and dimensions from those flat pictures.

The most common type is a structured light scanner, which projects a pattern of light onto an object, photographs how that pattern bends and warps across the surface, and calculates depth from the distortion. Another type, photogrammetry software, works differently — it takes multiple regular photographs of an object from different positions and stitches them together mathematically to build a 3D model. Your smartphone can do photogrammetry with the right app; professional scanners do it faster and more accurately.

A third type, laser triangulation, shoots a laser line at an object and measures where the reflection lands to determine distance. These scanners are common in manufacturing and quality control because they are fast and precise.

Key Takeaways

  • 3D scanners convert flat 2D images into digital 3D models by capturing photographs from multiple angles and using software to calculate depth and shape.
  • Structured light scanners project patterns onto objects and photograph the distortion to measure surface depth.
  • Photogrammetry software stitches together multiple regular photographs to build a 3D model, and works on smartphones as well as professional equipment.
  • Laser triangulation scanners measure laser reflections to determine distance and are used in manufacturing and quality control.
  • The output is a digital file — usually a point cloud or mesh — that can be rotated, measured, edited, or sent to a 3D printer.

How structured light scanning works

A structured light scanner projects a known pattern — usually a grid or series of stripes — onto the object you want to scan. A camera photographs that pattern as it sits on the surface. Where the surface is flat, the pattern stays regular. Where the surface curves or has depth, the pattern bends and stretches.

The scanner's software compares the projected pattern to the photographed pattern and calculates how far away each point on the surface must be to create that distortion. By doing this across thousands of points, it builds a complete 3D map. This method is fast — often taking seconds to minutes — and works well on objects with texture or detail. It struggles with shiny, transparent, or very dark surfaces because the camera cannot see the pattern clearly on those materials.

Photogrammetry: using regular photographs

Photogrammetry is simpler in concept but requires more photographs. You take 20 to 100 pictures of an object from different angles — walking around it, shooting from above, from the side, close up, far away. Photogrammetry software then identifies matching points across all those images (a corner, a texture detail, an edge) and uses the geometry of those matches to calculate where the camera was when each photo was taken and where the object's surface must be.

This method works on any object and any surface — shiny, dark, transparent, it does not matter. Your smartphone can run photogrammetry apps like Polycam or Scaniverse. Professional software like Agisoft Metashape or RealityCapture produces higher-quality models but costs money and requires more computing power. The trade-off is time: photogrammetry is slower than structured light scanning, sometimes taking hours to process a single object.

Laser triangulation for precision work

A laser triangulation scanner emits a thin laser line onto an object and measures where that line appears in a camera's view. The angle between the laser source, the point where the laser hits the object, and the camera creates a triangle. By measuring that triangle's angles and the known distance between the laser and camera, the software calculates the exact distance to that point on the surface.

The scanner moves the laser line across the object (or moves the object past the laser) to build a complete 3D map point by point. This method is extremely accurate — often to within fractions of a millimeter — which is why it is used in factories to check whether manufactured parts match their design specifications. It is slower than structured light and works best on matte surfaces, not shiny ones.

What the scanner outputs

All three scanning methods produce the same basic output: a point cloud or a mesh. A point cloud is millions of individual 3D coordinates — imagine a cloud of dots that outline the object's shape. A mesh connects those dots into triangles or polygons, creating a surface you can see and manipulate.

These files can be opened in 3D modeling software like Blender (free) or Fusion 360, edited to remove noise or fill gaps, and exported in formats like STL, OBJ, or PLY. You can measure distances and volumes, compare the scan to a design file to check for defects, or send the file to a 3D printer to create a physical copy.

Common uses for 3D scanning

Museums scan artifacts to create digital archives and 3D-printable replicas without handling fragile originals. Dentists scan teeth to design custom crowns and aligners. Manufacturers scan parts to verify they meet specifications or to reverse-engineer products. Video game developers scan real objects and people to create realistic digital models. Archaeologists scan dig sites and artifacts to document them precisely.

Smaller operations use smartphone photogrammetry to scan small objects for e-commerce photos, to document damage for insurance claims, or to create 3D models for hobbyist 3D printing. The barrier to entry has dropped significantly — a smartphone and free software can produce usable 3D models in minutes.

Choosing between scanning methods

If you need speed and are scanning textured objects indoors, a structured light scanner is the right choice. If you need accuracy and can work with matte surfaces, laser triangulation is best. If you need flexibility — scanning outdoors, handling reflective surfaces, or using equipment you already own — photogrammetry wins.

For most people, photogrammetry on a smartphone is the practical starting point. Apps like Polycam are free or low-cost, require no special hardware, and produce models good enough for visualization, 3D printing, or documentation. Professional scanners cost thousands of dollars and are worth the investment only if you scan regularly and need precision or speed that software cannot match.

Frequently Asked Questions

Can I scan something with my phone camera?

Yes, using photogrammetry apps. Take 30 to 50 photos of the object from different angles, upload them to an app like Polycam or Scaniverse, and the software builds a 3D model. Quality depends on lighting and how many photos you take, but the result is usually good enough for 3D printing or visualization.

What surfaces are hardest to scan?

Shiny, reflective, and transparent surfaces are difficult for structured light and laser scanners because the light does not reflect predictably. Photogrammetry handles these better because it relies on visible detail, not light reflection. Very dark or featureless surfaces are hard for all methods.

How accurate are 3D scans?

Photogrammetry on a smartphone is typically accurate to within 5 to 10 millimeters. Professional structured light scanners reach 0.1 to 1 millimeter. Laser triangulation can be accurate to 0.01 millimeters. Accuracy depends on the scanner, the object size, and the software used.

Can I edit a 3D scan after it is created?

Yes. Open the scan file in 3D software like Blender, Fusion 360, or Meshmixer. You can remove noise, fill holes, smooth surfaces, or combine multiple scans. Some software automates cleanup; other edits require manual work.

What file format should I use to share a 3D scan?

STL is the standard for 3D printing. OBJ and PLY are good for sharing models that will be viewed or edited in software. FBX is common in video game and animation work. Most 3D software can convert between formats.