US 11,989,870 B2
Method and assembly for detecting objects on systems
Josef Alois Birchbauer, Seiersberg (AT); Vlad Comanelea-Serban, Brasov (RO); and Olaf Kähler, Graz (AT)
Assigned to Siemens Energy Global GmbH & Co. KG, Bayern (DE)
Appl. No. 17/268,458
Filed by Siemens Energy Global GmbH & Co. KG, Bayern (DE)
PCT Filed Aug. 20, 2019, PCT No. PCT/EP2019/072269
§ 371(c)(1), (2) Date Feb. 13, 2021,
PCT Pub. No. WO2020/038944, PCT Pub. Date Feb. 27, 2020.
Claims priority of application No. 18190055 (EP), filed on Aug. 21, 2018.
Prior Publication US 2021/0174061 A1, Jun. 10, 2021
Int. Cl. G06T 17/00 (2006.01); G01S 17/89 (2020.01); G06T 7/00 (2017.01); G06V 20/10 (2022.01); G06V 20/17 (2022.01); G06V 20/64 (2022.01)
CPC G06T 7/0004 (2013.01) [G01S 17/89 (2013.01); G06T 17/00 (2013.01); G06V 20/17 (2022.01); G06V 20/176 (2022.01); G06V 20/64 (2022.01)] 13 Claims
OG exemplary drawing
 
1. A method for detecting objects on an installation, comprising:
providing a three-dimensional representation of the installation, wherein position and orientation of the representation and the installation are known, and
acquiring a first image and a second image of the installation, wherein the two images are recorded from different positions above the installation,
wherein for a plurality of sections of the installation, a comparison of the first and second images in each case is carried out making use of a parallax effect,
wherein in the case of a match between the images in a region surrounding the installation, an object is detected on the installation,
wherein the installation comprise one or more of an overhead line and a pipeline,
wherein the provided three-dimensional representation of the installation is used to restrict a search space for the installation or to assign the recognized object to a component of the installation in the three-dimensional representation, and
wherein the three-dimensional representation is captured as a three-dimensional point cloud (PCD), with the three-dimensional point cloud (PCD) being semantically segmented in order to restrict a search space for the installation in the three-dimensional point cloud (PCD).