Capturing and aligning three-dimensional scenes

A three-dimensional capture system, comprising:

• a rotatable stage that has mounted thereon, one or more three-dimensional capture devices configured to capture first images at various rotation points during a first rotation of the rotatable stage at a first location in an environment;
• a memory that stores computer executable instructions;
• and a processor, coupled to the memory, that facilitates execution of the executable instructions to perform operations, comprising:
• aligning the first images to one another using a pair-wise alignment process based on correspondences between position information determined for features included in respective first images of the first images, thereby generating a first set of aligned image data captured at the first location, wherein the position information is determined based on depth information respectively associated with the features;
• and
• sending the first set of aligned image data to a computation device, wherein based on the sending, the computation device is configured to perform a global alignment process that comprises aligning the first set of aligned image data with a second set of aligned image data in a common three-dimensional coordinate space to generate a three-dimensional reconstruction of the environment, wherein the second set of aligned image data comprises second images that were captured during a second rotation of the rotatable stage at a second location in the environment and aligned to one another using the pair-wise alignment process.

A method comprising:

• controlling, by a first device comprising a processor and memory, rotation of a rotatable stage comprising a plurality of three-dimensional capture devices mounted thereon;
• directing, by the first device, the plurality of three-dimensional capture devices to capture first images of an environment at different rotation points during a first rotation of the rotatable stage at a first location in the environment, wherein the first images are respectively associated with distance information for features respectively included therein;
• receiving, by the first device, the first images as they are captured during the first rotation;
• aligning, by the first device, the images to one another using a pair-wise alignment based on correspondences between position information determined for features included in respective first images of the first images, thereby generating a first set of aligned image data captured at the first location, wherein the position information is determined based on distance information respectively associated with the features;
• and
• sending, by the first device, the first set of aligned image data to a second device, wherein based on the sending, the second device is configured to generate a three-dimensional reconstruction of the environment by aligning the first set of aligned image data with a second set of aligned image data in a common three-dimensional coordinate space using a global alignment process, wherein the second set of aligned image data comprises second images that were captured during a second rotation of the rotatable stage at a second location in the environment and aligned to one another using the pair-wise alignment process.

A non-transitory computer-readable storage medium comprising computer-readable instructions that, in response to execution, cause a computing system to perform operations, comprising:

• controlling rotation of a three-dimensional capture device as the three-dimensional capture device captures first images of an environment at various rotation points relative to a first fixed location in the environment;
• receiving the first images as they are respectively captured;
• determining three-dimensional position information for points included in respective first images of the first images based on respective capture distances between the points and the three-dimensional capture device at times of capture of the respective first images;
• aligning the first images to one another using a pair-wise alignment process based on correspondences between the three-dimensional position information for the points, thereby generating a first set of aligned image data captured at the first fixed location;
• and
• generating a three-dimensional representation of the environment by further aligning the first set of aligned image data with a second set of aligned image data using a global alignment process, wherein the second set of aligned image data comprises second images that were captured during a second rotation of the rotatable stage at a second fixed location in the environment and aligned to one another using the pair-wise alignment process, and wherein the global alignment process comprises aligning the first set of aligned image data with the second set of aligned imaged data in a common three-dimensional coordinate space.