US 2017/0249776 A1Published application

Systems and methods for estimating pose of textureless objects

AssigneeCarnegie Mellon University

InventorsZhe Cao, Natasha Banerjee, Yaser Sheikh

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems and methods for estimating the pose of a textureless object are disclosed. A method to estimate a pose of a textureless object includes obtaining, by a processing device, a single image of the textureless object. The pose of the textureless object can be inferred from the single image. The method further includes generating, by the processing device, a three dimensional model of the textureless object from a plurality of viewpoints and a plurality of scales obtained from image data of the textureless object, matching, by the processing device, the single image with a discretized render of the three dimensional model via a multi-level illumination invariant tree structure to obtain an alignment of the single image with the discretized render, and estimating, by the processing device, the pose of the textureless object based on the alignment.

Claims20

Dependent claims

A system to output a pose of a textureless object, the system comprising:

a processing device;
and
a non-transitory, processor-readable storage medium, the non-transitory, processor-readable storage medium comprising one or more programming instructions that, when executed, cause the processing device to:
obtain a single image of the textureless object, wherein the pose of the textureless object is inferred from the single image;
generate a three dimensional model of the textureless object from a plurality of viewpoints and a plurality of scales obtained from image data of the textureless object;
extract a plurality of patches from each of a plurality of discretized renders of the three dimensional model, wherein the plurality of patches comprises a plurality of high-level patches and a plurality of low-level patches;
link each one of the plurality of high-level patches to at least one of the plurality of low-level patches to construct a multi-level illumination invariant tree structure;
match the single image with a discretized render of the three dimensional model via the multi-level illumination invariant tree structure to obtain an alignment of the single image with the discretized render;
and
output the pose of the textureless object based on the alignment.

7 more elements

A system to direct a robotic device based on a pose of a textureless object, the system comprising:

a processing device;
one or more imaging devices communicatively coupled to the processing device;
the robotic device communicatively coupled to the processing device;
a picking surface supporting the textureless object;
and
a non-transitory, processor-readable storage medium, the non-transitory, processor-readable storage medium comprising one or more programming instructions that, when executed, cause the processing device to:
obtain a single image of the textureless object from at least one of the one or more imaging devices, wherein the single image indicates the pose of the textureless object;
generate a three dimensional model of the textureless object from a plurality of viewpoints and a plurality of scales obtained from image data of the textureless object generated by the one or more imaging devices;
match the single image with a discretized render of the three dimensional model via a multi-level illumination invariant tree structure to obtain an alignment of the single image with the discretized render;
estimate the pose of the textureless object based on the alignment;
and
direct the robotic device to move and grasp the textureless object based on the estimated pose.

9 more elements

Dependencies are inferred from the claim text.