System and method for extracting and measuring shapes of objects having curved surfaces with a vision system
This invention provides a system and method that efficiently detects objects imaged using a 3D camera arrangement by referencing a cylindrical or spherical surface represented by a point cloud, and measures variant features of an extracted object including volume, height, and center of mass, bounding box, and other relevant metrics. The system and method, advantageously, operates directly on unorganized and un-ordered points, requiring neither a mesh/surface reconstruction nor voxel grid representation of object surfaces in a point cloud. Based upon a cylinder/sphere reference model, an acquired 3D point cloud is flattened. Object (blob) detection is carried out in the flattened 3D space, and objects are converted back to the 3D space to compute the features, which can include regions that differ from the regular shape of the cylinder/sphere. Downstream utilization devices and/or processes, such as part reject mechanism and/or robot manipulators can act on the identified feature data.
1. A system for detecting features relative to curved surfaces from a 3D point cloud imaged by a 3D vision system camera assembly comprising:
an extraction process that identifies a curved surface in 3D space and provides a reference model relative to the curved surface;
a flattening process that flattens the curved surface, including a feature thereof, into a flattened, planar shape in accordance with the reference model, wherein a portion of the curved surface is represented as a height image having values representative of distances of the points on the curved surface from a central element of the reference model, wherein the curved surface comprises one or more regions that are monotonic with respect to an outward radial direction from the central element of the reference model;
a feature location process that locates the feature on the planar shape represented in the height image by processing the planar shape produced from the flattening process with respect to the height image;
a mapping process that converts the feature located by processing the planar shape into the 3D space; and
a feature output process that reports, relative to the curved surface, the feature converted to the 3D space.
2. The system as set forth in claim 1 wherein the feature location process employs a blob analysis tool.
3. The system as set forth in claim 1 wherein the curved surface is a cylindrical surface or a spherical surface and the reference model, respectively, comprises a cylinder or a sphere.
4. The system as set forth in claim 1 wherein the flattening process flattens the curved surface along a surface tangent direction.
5. The system as set forth in claim 1 wherein the flattening process divides the planar surface, wherein 3D points of the curved surface reside, into a grid based upon orthogonal axes defining azimuth and height on a cylinder or azimuth and tilt on a sphere.
6. The system as set forth in claim 1 , further comprising, a feature identification process that identifies features differing from a normal surface geometry of the curved surface.
7. The system as set forth in claim 6 wherein a differing feature comprises an object defect including at least one of a deformation, a protrusion, a dent, a dimple, or a scratch.
8. The system as set forth in claim 1 , further comprising, a reference surface refinement process that adapts the reference model to 3D points representing the planar surface.
9. The system as set forth in claim 8 , wherein the reference surface refinement process adapts the reference model based on an estimated noise threshold derived from the height image.
10. The system as set forth in claim 9 , wherein the estimated noise threshold is arranged to distinguish foreground pixels from background pixels of the height image.
11. A method for detecting features relative to curved surfaces from a 3D point cloud imaged by a 3D vision system camera assembly comprising:
identifying a curved surface in 3D space and providing a reference model relative to the curved surface;
flattening the curved surface into a flattened, planar shape in accordance with the reference model, wherein a portion of the curved surface is represented as a height image having values representative of distances of points on the curved surface from a central element of the reference model, wherein the curved surface comprises one or more regions that are monotonic with respect to an outward radial direction from the central element of the reference model;
locating a feature on the planar shape represented in the height image by processing the planar shape produced from the flattening process with respect to the height image;
mapping the located feature into the 3D space; and
reporting the located feature relative to the curved surface.
12. The method as set forth in claim 11 wherein the locating of the feature employs a blob analysis tool.
13. The method as set forth in claim 11 wherein the curved surface is a cylindrical surface or a spherical surface and the reference model, respectively, comprises a cylinder or a sphere.
14. The method as set forth in claim 11 wherein the flattening flattens the curved surface along a surface tangent direction.
15. The method as set forth in claim 11 wherein the flattening divides the planar surface, wherein 3D points of the curved surface reside, into a grid based upon orthogonal axes defining azimuth and height on a cylinder or azimuth and tilt on a sphere.
16. The method as set forth in claim 11 , further comprising, identifying a feature differing from a normal surface geometry of the curved surface.
17. The method as set forth in claim 16 wherein the differing feature comprises an object defect including at least one of a deformation, a protrusion, a dent, a dimple, or a scratch.
18. The method of claim 11 , wherein the reporting comprises outputting a 3D graphical representation of the located feature relative to the curved surface.
19. A system for detecting object features within an industrial process, the process comprising:
a 3D vision system camera including an image sensor and a field of view (FOV), and configured to acquire a 3D point cloud representative of an object within the FOV; and
at least one processor in communication with the 3D vision camera and configured to:
receive the 3D point cloud representative of the object;
identify, from the 3D point cloud, a curved surface in 3D space;
provide a reference model relative to the curved surface;
flatten the curved surface, including a feature thereof, into a flattened, planar shape in accordance with the reference model, wherein a portion of the curved surface is represented as a height image having values representative of distances of points on the curved surface from a central element of the reference model, wherein the curved surface comprises one or more regions that are monotonic with respect to an outward radial direction from the central element of the reference model;
locate the feature on the planar shape within the height image by processing the planar shape with respect to the height image;
convert the located feature in to the 3D space; and
output the located feature relative to the curved surface within the 3D space.
20. The method of claim 18 , wherein the 3D graphical representation includes at least one of a highlighted region or a bounding box corresponding to the located feature.