Back to results
US 12,722,278 B2Granted
Mobile robot with improved accuracy and generating three-dimensional depth map
AssigneePixart Imaging Inc.
InventorsGuo-Zhen Wang, Tse-En Peng
Life of this patent
Estimating the term…
Abstract
There is provided a mobile robot that performs the de-flickering and different auto exposures in a pixel array in the range estimation to be adaptable to different operating scenarios, and constructs a three-dimensional depth map to lower the cost.
Claims20
A mobile robot, comprising:
- a light source, configured to project a linear light section of infrared light toward a moving direction of the mobile robot;
- an optical sensor, comprising a plurality of infrared pixels and a plurality of non-infrared pixels, and configured to capture an image frame toward the moving direction;
- and
- a processor, embedded with a machine learning algorithm and coupled to the optical sensor, and configured to
- divide the image frame into a first sub-frame, associated with the plurality of non-infrared pixels, and a second sub-frame, associated with the plurality of infrared pixels,
- calculate, using the machine learning algorithm, relative depths of obstacles in the first sub-frame,
- calculate absolute depths of the obstacles in the second sub-frame, and
- construct a three-dimensional depth map by modifying each of the relative depths using a corresponding absolute depth among the calculated absolute depths.
5 more elements
A mobile robot, comprising:
- a light source, configured to project a longitudinal light section toward a moving direction of the mobile robot, and the light source being turned on and turned off alternatively;
- an optical sensor, configured to sequentially capture a first dark image frame, a bright image frame and a second dark image frame toward the moving direction;
- and
- a processor, coupled to the optical sensor, and configured to
- calculate a first differential image frame between the first dark image frame and the second dark image frame to determine at least one flicker region,
- calculate a second differential image frame between the bright image frame and one of the first dark image frame and the second dark image frame, and
- use the at least one flicker region as at least one flicker mask in the second differential image frame.
4 more elements
A mobile robot, comprising:
- a light source, configured to project a longitudinal light section toward a moving direction of the mobile robot;
- an optical sensor, comprising a pixel array divided into an upper part pixels and a lower part pixels, and the pixel array being configured to capture an image frame toward the moving direction;
- and
- a processor, coupled to the optical sensor, and configured to
- control the upper part pixels to perform a first auto exposure, and
- control the lower part pixels to perform a second auto exposure individual from the first auto exposure.
3 more elements
Dependencies are inferred from the claim text.