Shape measuring apparatus, shape measuring method, non-transitory computer readable medium storing program
AssigneeJVCKENWOOD Corporation
InventorTakashi Himukashi
Life of this patent
Estimating the term…
Abstract
Provided is a shape measuring apparatus, a shape measuring method, and a program that can efficiently measure a three-dimensional shape of an object. The shape measuring apparatus includes an imaging unit, a distance image sensor unit, and a processing apparatus. The imaging unit acquires a two-dimensional image of an object over a plurality of frames. The distance image sensor unit acquires a distance image of the object over the plurality of frames. The processing apparatus detects a feature point of the object from the distance image of each of the plurality of frames and writes the detected feature point in a coordinate system set for the object.
Claims12
A shape measuring apparatus comprising:
- an imaging unit configured to acquire a two-dimensional image of an object over a plurality of frames, the object being held by a hand of a user of the shape measuring apparatus;
- a distance image sensor unit configured to acquire a distance image of the object over the plurality of frames, the distance image of the object being three-dimensional point cloud data including coordinates of pixels on a two-dimensional plane and the distances at the coordinates;
- a processing apparatus configured to detect a feature point of the object from the distance image of each of the plurality of frames that has been acquired by the user moving the object, to write the detected feature point in a coordinate system set for the object, and to delete data of the hand of the user from the distance image of the object;
- and
- a division processing unit configured to divide a three-dimensional model obtained from the distance image, wherein
- the processing apparatus deletes the feature point that has been detected a small number of times as the hand of the user from the data that has been acquired by performing a plurality of detections while the user changes a holding position of the object, and
- the division processing unit obtains an area of a cross-section of the three-dimensional model sequentially selected at a predetermined pitch along a predetermined axis and divides the three-dimensional model by either of a currently selected cross-section or an adjacent cross-section selected just before the currently selected cross-section when an area difference between the currently selected cross-section and the adjacent cross-section is larger than a predetermined threshold value
- wherein at least one predetermined threshold value in three predetermined threshold values respectively corresponding to the three axes is different from the other predetermined threshold values.
A shape measuring method comprising:
- causing an imaging unit to acquire a two-dimensional image of an object over a plurality of frames, the object being held by a hand of a user that performs a measurement;
- causing a distance image sensor unit to acquire a distance image of the object over the plurality of frames, the distance image of the object being three-dimensional point cloud data including coordinates of pixels on a two-dimensional plane and the distances at the coordinates;
- causing a processing apparatus to:
- detect a feature point of the object from the distance image of each of the plurality of frames that has been acquired by the user moving the object;
- write the detected feature point in a coordinate system set for the object;
- and
- delete data of the hand of the user from the distance image of the object;
- and
- causing a division processing unit to divide a three-dimensional model obtained from the distance image, wherein
- the processing apparatus deletes the feature point that has been detected a small number of times as the hand of the user from the data that has been acquired by performing a plurality of detections while the user changes a holding position of the object, and
- the division processing unit obtains an area of a cross-section of the three-dimensional model sequentially selected at a predetermined pitch along a predetermined axis and divides the three-dimensional model by either of a currently selected cross-section or an adjacent cross-section selected just before the currently selected cross-section when an area difference between the currently selected cross-section and the adjacent cross-section is larger than a predetermined threshold value
- wherein at least one predetermined threshold value in three predetermined threshold values respectively corresponding to the three axes is different from the other predetermined threshold values.
A non-transitory computer readable medium storing a shape measuring program that causes a computer to execute:
- acquiring a two-dimensional image of an object over a plurality of frames, the object being held by a hand of a user that performs a measurement;
- acquiring a distance image of the object over the plurality of frames, the distance image of the object being three-dimensional point cloud data including coordinates of pixels on a two-dimensional plane and the distances at the coordinates;
- detecting a feature point of the object from the distance image of each of the plurality of frames that has been acquired by the user moving the object;
- writing the detected feature point in a coordinate system set for the object;
- deleting data of the hand of the user from the distance image of the object;
- dividing a three-dimensional model obtained from the distance image, wherein
- the feature point that has been detected a small number of times is deleted as the hand of the user from the data that has been acquired by performing a plurality of detections while the user changes a holding position of the object, and
- an area of a cross-section of the three-dimensional model sequentially selected at a predetermined pitch along a predetermined axis is obtained and the three-dimensional model is divided by either of a currently selected cross-section or an adjacent cross-section selected just before the currently selected cross-section when an area difference between the currently selected cross-section and the adjacent cross-section is larger than a predetermined threshold value
- wherein at least one predetermined threshold value in three predetermined threshold values respectively corresponding to the three axes is different from the other predetermined threshold values.
Dependencies are inferred from the claim text.