US 12,260,655 B2
Method for detection of three-dimensional objects and electronic device
Chih-Te Lu, New Taipei (TW); Chieh Lee, New Taipei (TW); and Chin-Pin Kuo, New Taipei (TW)
Assigned to HON HAI PRECISION INDUSTRY CO., LTD., New Taipei (TW)
Filed by HON HAI PRECISION INDUSTRY CO., LTD., New Taipei (TW)
Filed on Jun. 30, 2022, as Appl. No. 17/854,301.
Claims priority of application No. 202210616195.7 (CN), filed on May 31, 2022.
Prior Publication US 2023/0386230 A1, Nov. 30, 2023
Int. Cl. G06V 20/64 (2022.01); G06T 7/70 (2017.01); G06V 20/58 (2022.01)
CPC G06V 20/64 (2022.01) [G06T 7/70 (2017.01); G06V 20/58 (2022.01); G06T 2200/04 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method for detection of three-dimensional (3D) objects comprising:
obtaining detection images by a camera;
inputting the detection images into a trained object detection model, and determining object categories, two-dimensional (2D) bounding boxes and rotation angles of objects in the detection images by using the trained object detection model;
according to the object categories, determining object models of the objects and 3D bounding boxes of the object models from a 3D object model library;
determining a distance from the camera to the object models according to a size of the 2D bounding boxes of the objects, image information of the detection images, and a focal length of the camera, by: calculating the distance from the camera to the object models according to a width and/or a length of the 2D bounding boxes of the objects, the focal length of the camera, a resolution of the detection images, and a pixel width of the detection images;
determining a position of the object models in a 3D space according to the rotation angles of the objects, the distance from the camera to the object models and the 3D bounding boxes, and taking the position of the object models in the 3D space as the position of the objects in the 3D space.