Photogrammetry is a technique that creates a 3D model from multiple photos that overlap by at least 70%. Differences in perspective allow the computer to calculate the physical geometry, with millimeter precision in some cases.
NOTE: I try to include a tape measure with big numbers in my scans so I can rescale the model to reality
LiDAR scanners like those on high-end iPhones and iPads use reflected light to capture 3D models with high accuracy. But the sensor is limited to around 1 cm, so LiDAR scans often lack detail that you can get with photogrammetry:

LiDAR

Photogrammetry - more detail

Photogrammetry - close to object, even more detail
Photogrammetry using still photos is better than LiDAR
| LiDAR | Experience |
|---|---|
| + | Measurements are accurate right away and do not need to be rescaled |
| + | Better at capturing reflecting or transparent surfaces than photogrammetry |
| - | Model quality is often poor compared with photogrammetry |
| - | Model experiences "drift" and is not accurate enough (a gimbal helps reduce drift) |
| Photogrammetry | Expérience |
|---|---|
| + | This technique with a simple cell phone can now outperform laser scanning |
| + | The quality and detail is better, especially for close objects |
| - | Cannot handle transparent or reflective surfaces like windows and water |
| - | The model must be rescaled. Not a big deal if you have also taken tape measurements and know how to do it. I always capture with a tape measure in the model so I can double-check that a model is accurately scaled. |