Image-stitching for dimensioning
First Claim
1. A method, comprising:
- projecting, using a pattern projector, a light pattern into a field-of-view of a range camera of a dimensioning system, the field-of-view comprising a portion of an object;
capturing, using the range camera of the dimensioning system, a range image of the field-of-view, wherein each pixel of the range image represents a distance from the range camera to a respective point in the range camera'"'"'s field-of-view, wherein the range image is created by determining a range value for each point of light in the projected light pattern to generate a point-cloud image and interpolating range values between points of light in the point-cloud image;
moving the dimensioning system and/or the object so that there is relative movement between the dimensioning system and the object, and the range camera'"'"'s field-of-view contains a different portion of the object;
repeating the capturing and the moving until a plurality of range images have been captured, wherein in each range image of the plurality of range images, each pixel of the range image represents a distance from the range camera to a respective point in the range camera'"'"'s field-of-view;
gathering information as a spatial relationship between the range camera and the object is changed, and the information is selected from the group consisting of tracking information and mapping information;
combining the plurality of range images to create a composite range-image by image-stitching the plurality of range images using the information gathered as the spatial relationship between the range camera and the object is changed; and
dimensioning the object using the composite range-image, wherein dimensioning the object using the composite range-image comprises creating a minimum bounding box that surrounds the object; and
returning the dimensions of the minimum bounding box as dimensions of the object.
0 Assignments
0 Petitions
Accused Products
Abstract
Dimensioning systems may automate or assist with determining the physical dimensions of an object without the need for a manual measurement. A dimensioning system may project a light pattern onto the object, capture an image of the reflected pattern, and observe changes in the imaged pattern to obtain a range image, which contains 3D information corresponding to the object. Then, using the range image, the dimensioning system may calculate the dimensions of the object. In some cases, a single range image does not contain 3D data sufficient for dimensioning the object. To mitigate or solve this problem, the present invention embraces capturing a plurality of range images from different perspectives, and then combining the range images (e.g., using image-stitching) to form a composite range-image, which can be used to determine the object'"'"'s dimensions.
1125 Citations
17 Claims
-
1. A method, comprising:
-
projecting, using a pattern projector, a light pattern into a field-of-view of a range camera of a dimensioning system, the field-of-view comprising a portion of an object; capturing, using the range camera of the dimensioning system, a range image of the field-of-view, wherein each pixel of the range image represents a distance from the range camera to a respective point in the range camera'"'"'s field-of-view, wherein the range image is created by determining a range value for each point of light in the projected light pattern to generate a point-cloud image and interpolating range values between points of light in the point-cloud image; moving the dimensioning system and/or the object so that there is relative movement between the dimensioning system and the object, and the range camera'"'"'s field-of-view contains a different portion of the object; repeating the capturing and the moving until a plurality of range images have been captured, wherein in each range image of the plurality of range images, each pixel of the range image represents a distance from the range camera to a respective point in the range camera'"'"'s field-of-view; gathering information as a spatial relationship between the range camera and the object is changed, and the information is selected from the group consisting of tracking information and mapping information; combining the plurality of range images to create a composite range-image by image-stitching the plurality of range images using the information gathered as the spatial relationship between the range camera and the object is changed; and dimensioning the object using the composite range-image, wherein dimensioning the object using the composite range-image comprises creating a minimum bounding box that surrounds the object; and returning the dimensions of the minimum bounding box as dimensions of the object. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11)
-
-
12. A dimensioning system, comprising:
-
a pattern projector configured to project a light pattern onto an object; a range camera having a field of view and configured to (i) capture an image of a reflected light-pattern in the field of view, (ii) generate 3D data from the reflected light-pattern, and (iii) create a range image using the 3D data by determining a range value for each point of light in the projected light pattern to generate a point-cloud image and interpolating range values between points of light in the point-cloud image, wherein the pattern projector and the range camera are spatially offset; at least one device configured to move the dimensioning system and/or the object so that there is relative movement between the dimensioning system and the object, and the range camera'"'"'s field-of-view contains a different portion of the object; and a processor communicatively coupled to the pattern projector and the range camera, wherein the processor is configured by software to; trigger the range camera to capture a plurality of range images, wherein in each range image of the plurality of range images, each pixel of the range image represents a distance from the range camera to a respective point in the range camera'"'"'s field-of-view; gather information as the spatial relationship between the range camera and the object is changed, and the information is selected from the group consisting of tracking information and mapping information; combine the plurality of range images to create a composite range-image by image-stitching the plurality of range images using the information gathered as the spatial relationship between the range camera and the object is changed; and dimension the object using the composite range-image, wherein dimensioning the object using the composite range-image comprises creating a minimum bounding box that surrounds the object; and return the dimensions of the minimum bounding box as dimensions of the object. - View Dependent Claims (13, 14, 15, 16, 17)
-
Specification