Gaussian Smoothing

The Gaussian filter performs smoothing of points within a small local region based on 3D distances. This can be used to suppress sparse noise, correct outliers, and align points to a grid. Sigma parameter defines the aggressiveness of the filter. The higher the sigma, the more aggressive smoothing. Note that the Gaussian filter can, in many cases, correct outliers, but high sigma will also smooth out edges.

Note

Gaussian filter adds the following processing time per 3D capture:

Expected median (±stddev) in ms

Filter	Zivid One+			Zivid Two
	Intel UHD 750	Intel UHD G1	NVIDIA 3070	Intel UHD 750	Intel UHD G1	NVIDIA 3070
	High-end [1]	Low-end [2]	High-end [3]	High-end [1]	Low-end [2]	High-end [3]
Smoothing	7 (±2) ms	14 (±2) ms	2 (±0.8) ms	6 (±2) ms	13 (±445) ms	2 (±0.8) ms

[1] (1,2)

High-end machine with GPU: Intel UHD Graphics 750 (ID:0x4C8A) and CPU: 11th Gen Intel(R) Core(TM) i9-11900K @ 3.50GHz, 10GbE

[2] (1,2)

Low-end machine with GPU: Intel UHD Graphics G1 (ID:0x8A56) and CPU: Intel(R) Core(TM) i3-1005G1 CPU @ 1.20GHz, 1GbE

[3] (1,2)

High-end machine with GPU: NVIDIA GeForce RTX 3070 and CPU: 11th Gen Intel(R) Core(TM) i9-11900K @ 3.50GHz, 10GbE

When do you use the Gaussian smoothing?

Because the Gaussian filter performs smoothing, it can greatly improve the absolute noise in a point cloud. This can improve the performance of many vision algorithms, such as detection and matching algorithms. Even though the relative noise from one part to another is the same, the absolute noise may be reduced, making the parts more similar as a whole.

The Gaussian filter can also be good to use when assessing large-scale effects, such as the flatness of a plane. This is because it suppresses the impact of outliers that may cause the applied algorithm to perform too pessimistically.

Version History

SDK	Changes
1.3.0	Gaussian filter API is added.
2.9.0	Reduce filter processing time and improve edge preservation capabilities.