US 12,430,802 B2
Calibrating a camera for mapping image pixels to grid points in a storage system
Xiaokai Cui, Hatfield (GB); Christopher Pearman, Hatfield (GB); and Jonathan Schuchart, Hatfield (GB)
Assigned to Ocado Innovation Limited, Hatfield (GB)
Appl. No. 18/713,075
Filed by Ocado Innovation Limited, Hatfield (GB)
PCT Filed Nov. 25, 2022, PCT No. PCT/EP2022/083366
§ 371(c)(1), (2) Date May 23, 2024,
PCT Pub. No. WO2023/094635, PCT Pub. Date Jun. 1, 2023.
Claims priority of application No. 2117102 (GB), filed on Nov. 26, 2021.
Prior Publication US 2025/0037310 A1, Jan. 30, 2025
Int. Cl. G06T 7/80 (2017.01)
CPC G06T 7/80 (2017.01) [G06T 2207/20084 (2013.01); G06T 2207/30204 (2013.01)] 25 Claims
OG exemplary drawing
 
1. A method of calibrating a wide-angle or ultra wide-angle camera disposed above a grid of a storage system, wherein the grid is formed by a first set of parallel tracks extending in an X-direction, and a second set of parallel tracks extending in a Y-direction transverse to the first set in a substantially horizontal plane,
the grid comprising a plurality of grid spaces and arranged to support one or more transport devices, selectively movable in at least one of the X-direction or Y-direction on the tracks, for handling a container stacked beneath the tracks within a footprint of a single grid space,
the method comprising:
obtaining an image of a grid section of the grid captured by the camera;
obtaining initial values of a plurality of parameters corresponding to the camera, the plurality of parameters including a focal length of the camera, a translational vector representative of a position of the camera above the grid section, and a rotational vector representative of a tilt and rotation of the camera;
processing the image, using a neural network trained to detect/predict the first and second sets of parallel tracks in images of grid sections captured by wide-angle or ultra wide-angle cameras, to generate a model of the first and second sets of parallel tracks as captured in the image of the grid section;
mapping selected pixels in the model to corresponding points on the grid using a mapping based on the plurality of parameters, wherein the initial values are used as inputs to the mapping;
determining an error function based on a discrepancy between grid coordinates of the points determined by the mapping and known grid coordinates;
updating the initial values to updated values of the plurality of parameters based on the error function; and
storing the updated values of the plurality of parameters for mapping pixels in images of the grid section captured by the camera to corresponding points on the grid of the storage system via the mapping.