Three-dimiensional point cloud generation using machine learning
US-2022311987-A1 · Sep 29, 2022 · US
US12276402B2 · US · B2
| Field | Value |
|---|---|
| Publication number | US-12276402-B2 |
| Application number | US-202217737752-A |
| Country | US |
| Kind code | B2 |
| Filing date | May 5, 2022 |
| Priority date | May 6, 2021 |
| Publication date | Apr 15, 2025 |
| Grant date | Apr 15, 2025 |
A practical reading order for non-experts. Skip the full description unless you need deep technical detail.
What the patent document calls the invention.
A short plain-language summary of the technical disclosure.
Who owns or filed the patent and who is credited as inventor.
Filing, priority, publication, and grant dates set the timeline.
The legal scope of protection — read this for what is actually claimed.
Technology tags used to group this patent with similar filings.
Prior art links and similar publications in this corpus.
Official abstract text for this publication.
A computer-vision system is equipped with a first camera and a second camera spaced apart from each other and enables acquisition of a first image and a second image of a light beam that propagates from a light-beam source towards a target position of the light beam. The first image and the second image of the beam differ from each other as a result of the first camera and the second camera being spaced apart from each other. The system produces a disparity map of the first image and second image of the beam to be projected in a three-dimensional point cloud starting from the disparity map according to point-cloud-projection calibration data that set in relation disparity maps for reference images with respective three-dimensional point clouds.
Opening claim text (preview).
What is claimed is: 1. A method comprising: acquiring via a computer-vision system equipped with a first camera and a second camera spaced apart from each other, a first image and a second image of a light beam that propagates from a light-beam source towards a target position of the light beam, wherein the first image and the second image of the light beam differ from each other as a result of the first camera and the second camera being spaced apart from each other, producing a disparity map of said first image and said second image of the beam, projecting a three-dimensional point cloud starting from said disparity map of said first image and said second image of the beam, clustering the points in said three-dimensional point cloud to produce a geometric line that extends between said light-beam source and said target position of the light beam, applying a variation of orientation on said light-beam source, and identifying a first end of said geometric line between said light-beam source and said target position of the light beam subject to displacement in response to said variation of orientation and a second end of said geometric line between said light-beam source and said target position of the light beam exempt from displacement in response to said variation of orientation, wherein said first end and said second end identify, with respect to said geometric line, the location of said target position of the light beam and the location of said light-beam source, respectively. 2. The method of claim 1 , further comprising: storing in the computer-vision system point cloud projection calibration data that relates disparity maps for reference images acquired through said first camera and said second camera with respective three-dimensional point clouds, and projecting said three-dimensional point cloud starting from said disparity map of said first image and said second image of the beam as a function of said point cloud projection calibration data stored in the computer vision system. 3. The method of claim 1 , further comprising applying to said first image and to said second image of the beam, before producing said disparity map, processing comprising at least one of contrast enhancement and thresholding. 4. The method of claim 1 , further comprising applying to said first image and to said second image of the beam, before producing said disparity map, edge-search processing. 5. The method of claim 1 , further comprising applying to said first image and to said second image of the beam, before producing said disparity map, line-search processing. 6. The method of claim 1 , further comprising: locating a first target position of the light beam starting from said light-beam source as a function of said geometric line that extends between said light-beam source and said first target position of the light beam, comparing said first target position of the light beam starting from said light-beam source with a second predetermined target position and detecting the offset between said first target position and said second target position, and adjusting the orientation of said light beam that propagates from said light-beam source as a function of the offset detected between said first target position and said second target position, wherein, in response to said adjustment, said light beam propagates from said light-beam source towards said second target position. 7. The method of claim 6 , wherein said adjustment comprises an open-loop adjustment as a function of the offset detected between said first target position and said second target position. 8. The method of claim 6 , wherein said second target position is user identifiable as one of: a position propagation of said light beam from said light-beam source, or a target straight line to make said light beam parallel to. 9. The method of claim 1 , applied to a lighting system comprising a plurality of light-beam sources that produce respective light beams, the method comprising selectively activating said sources in said plurality of light-beam sources during respective activation intervals, wherein, during the respective activation interval, one source in the plurality of light-beam sources is activated to produce a respective light beam while the other sources in the plurality of light-beam sources are deactivated, wherein the method comprises carrying out, during said respective activation interval for a respective light beam produced by the light-beam source activated in said respective activation interval, said steps of acquiring a first image and a second image of the beam, producing said disparity map of said first image and said second image of the beam, projecting said three-dimensional point cloud and clustering the points in said three-dimensional point cloud to produce a geometric line extending between the source in the plurality of light-beam sources activated to produce a respective light beam and said target position of the light beam. 10. A lighting system comprising: at least one light-beam source configured to produce a light beam that propagates from said at least one light-beam source towards a target position of the light beam, and a computer-vision system equipped with a first camera and a second camera spaced apart from each other, wherein the computer-vision system is configured to perform steps of acquiring a first image and a second image of the beam, producing a disparity map of said first image and said second image of the beam, projecting a three-dimensional point cloud and clustering the points in said three-dimensional point cloud to produce a geometric line, applying a variation of orientation on said light-beam source, and identifying a first end of said geometric line between said light-beam source and said target position of the light beam subject to displacement in response to said variation of orientation and a second end of said geometric line between said light-beam source and said target position of the light beam exempt from displacement in response to said variation of orientation, wherein said first end and said second end identify, with respect to said geometric line, the location of said target position of the light beam and the location of said light-beam source, respectively. 11. A method comprising: acquiring via a computer-vision system equipped with a first camera and a second camera spaced apart from each other, a first image and a second image of a light beam that propagates from a light-beam source towards a target position of the light beam, wherein the first image and the second image of the light beam differ from each other as a result of the first camera and the second camera being spaced apart from each other, producing a disparity map of said first image and said second image of the beam, projecting a three-dimensional point cloud starting from said disparity map of said first image and said second image of the beam, clustering the points in said three-dimensional point cloud to produce a geometric line that extends between said light-beam source and said target position of the light beam, locating a first target position of the light beam starting from said light-beam source as a function of said geometric line that extends between said light-beam source and said first target position of the light beam, comparing said first target position of the light beam starting from said light-beam source with a second predetermined target position and detecting the offset between said first target position and said second target position, and adjusting the orientation of said light beam that propagates from said light-beam source as a function of the offset detected between said
using digital multiplexed [DMX] communication protocols · CPC title
by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation · CPC title
Range image; Depth image; 3D point clouds · CPC title
for theatres, stages or film studios · CPC title
by means of an image recording device, e.g. a camera · CPC title
Related publications grouped by family.
Answers are generated from the same data shown on this page.