Embedded Files
A LiDAR sensor is available on the iPhone 12 Pro and later Pro models. This sensor allows for non-contact depth sensing at distances up to 5 meters, but practically and most accurately between 20 cm and 2 m. At close range, displacements can be measured with an uncertainty of ±1 mm, but vibrations as small as 0.1 mm are detectable.
Apps
Apps
PhyPhox allows for patch selection with aggregation of depth values according to an average, weighted, or closest approach. It is very effective for single-point vibration measurements.
Physics Toolbox Sensor Suite has features similar to PhyPhox, but with lower precision and fewer post-processing tools.
Record 3D has the ability to record 3D videos, capturing dense point clouds at 60 frames per second. The potential of this tool to measure full-field displacements is still being explored.
StrayScanner can produce a 256x192 depth map, the rawest form available from the Apple AR Kit, at various frame rates (60, 30, 15, 5, and 1 frames per second).
Example
Example
I measured the natural frequency of a steel cantilever beam to be 2.77 Hz using the PhyPhox app and front-facing LiDAR on an iPhone 13 Pro. After collecting the time history, I picked points at the top of the peaks of 5 cycles to find a time of 1.8029 seconds. Dividing by 5 yields a natural period of 0.3606 seconds, or a natural frequency of 2.77 Hz.
I measured the natural frequency of the same steel cantilever beam to be 2.78 Hz using the PhyPhox app and rear-facing LiDAR on an iPhone 13 Pro. Selecting an appropriate patch and the "closest" aggregation mode ensures that the values associated with the beam in the foreground are used. The duration of 7 cycles was measured as 2.520 seconds, meaning a natural period of 0.360 seconds, or a natural frequency of 2.78 Hz. This is within the range of 2.70 to 2.79 Hz measured by the various methods described on this site.
Verification of LiDAR With High-Precision System
Verification of LiDAR With High-Precision System
I placed an iPhone 13 Pro running PhyPhox at the tip of a cantilevered aluminum sign. I measured vibrations from the most sensitive 10-V/g accelerometer I have available. Analyzing with the Vibration Analysis app yielded 2.358 Hz. The phone's on-board accelerometer registered 2.360 Hz. The LiDAR measurement indicated that 11 cycles took 7.127 seconds, for a period of 0.423 seconds, or a frequency of 2.365 Hz. Each of these measurements is within 0.3% of the other, which is very good agreement and demonstrates the accuracy of these phone-based methods.
Noise Characterization
Noise Characterization
The noise inherent in the LiDAR signal can be characterized by recording a static measurement. Setting the phone approximately 0.25 m from a fixed surface and collecting a 4.3-minute measurement resulted in the following results, indicating a normal distribution of values around a mean value of 250.18 mm and a range of 3 mm. Thus, at close range, results are within ±1 mm.
There is an excellent visualization of the LiDAR array in this excellent video by the Action Lab:
Page updated
Google Sites
Report abuse