Systems and methods for identifying a discontinuity in the boundary of an object in an image
Machine vision is used to identify a discontinuity in the boundary of an object in an image. An image of one or more objects is captured. One or more skeletons of the one or more objects are calculated. One or more boundaries of the one or more objects are calculated. A plurality of radial lines is extended from a spine point of a skeleton to the one or more boundaries. Each radial line intersects a boundary at a radial endpoint producing a plurality of radial endpoints. For each radial endpoint an expected radial endpoint is calculated based on two or more neighboring radial endpoints. If the difference between the radial endpoint and its expected radial endpoint exceeds a threshold, a radial line including the radial endpoint is identified as a discontinuity in a boundary of an object.
1. A system for identifying a discontinuity in the boundary of an object in an image, comprising:
a light source that illuminates one or more objects;
an imaging device that captures an image of the one or more objects;
a processor in communication with the light source and the imaging device that
calculates one or more skeletons of the one or more objects in the captured image,
calculates one or more boundaries of the one or more objects in the captured image,
extends a plurality of radial lines from a spine point of a skeleton of the one or more skeletons to the one or more boundaries, wherein each radial line of the plurality of radial lines intersects a boundary of the one or more boundaries at a radial endpoint producing a plurality of radial endpoints,
calculates for each radial endpoint of the plurality of radial endpoints an expected radial endpoint based on two or more neighboring radial endpoints, and
if the difference between the each radial endpoint and its expected radial endpoint exceeds a threshold, identifies a radial line including the each radial endpoint as a discontinuity in a boundary of an object of the one or more objects.
2. The system of claim 1 , wherein the discontinuity identifies a merged or occluded area of the object.
3. The system of claim 1 , wherein the processor calculates an expected radial endpoint based on two or more neighboring radial endpoints using a curve fitting algorithm.
4. The system of claim 3 , wherein the curve fitting algorithm is chosen based on application domain knowledge.
5. The system of claim 3 , wherein the curve fitting algorithm is chosen based on sensor information.
6. The system of claim 1 , wherein the processor further removes the discontinuity in the boundary of the object.
7. The system of claim 6 , wherein the processor removes the discontinuity by recalculating the expected radial endpoint of the radial line based on two or more next nearest neighboring radial endpoints.
8. The system of claim 6 , wherein the processor removes the discontinuity by adding one or more additional radial lines to the plurality of radial lines that extend from the spine point and recalculating the expected radial endpoint of the radial line based on radial endpoints of one or more additional radial lines.
9. The system of claim 6 , wherein the processor removes the discontinuity by removing the radial line from the plurality of radial lines.
10. The system of claim 6 , wherein the processor calculates shape features of the object based on the plurality of radial endpoints.
11. The system of claim 10 , wherein the processor calculates the shape features of the object based on the plurality of radial endpoints using a curve fitting algorithm.
12. The system of claim 3 , wherein the curve fitting algorithm is chosen based on application domain knowledge.
13. The system of claim 3 , wherein the curve fitting algorithm is chosen based on sensor information.
14. A method for identifying a discontinuity in the boundary of an object in an image, comprising:
illuminating one or more objects using a light source;
capturing an image of the one or more objects using an imaging device;
calculating one or more skeletons of the one or more objects in the captured image using a processor;
calculating one or more boundaries of the one or more objects in the captured image using the processor;
extending a plurality of radial lines from a spine point of a skeleton of the one or more skeletons to the one or more boundaries using the processor, wherein each radial line of the plurality of radial lines intersects a boundary of the one or more boundaries at a radial endpoint producing a plurality of radial endpoints;
calculating for each radial endpoint of the plurality of radial endpoints an expected radial endpoint based on two or more neighboring radial endpoints using the processor; and
if the difference between the each radial endpoint and its expected radial endpoint exceeds a threshold, identifying a radial line including the each radial endpoint as a discontinuity in a boundary of an object of the one or more objects using the processor.
15. The method of claim 14 , wherein further comprising removing the discontinuity in the boundary of the object using the processor.
16. The method of claim 15 , wherein removing the discontinuity comprises recalculating the expected radial endpoint of the radial line based on two or more next nearest neighboring radial endpoints.
17. The method of claim 15 , wherein removing the discontinuity comprises adding one or more additional radial lines to the plurality of radial lines that extend from the spine point and recalculating the expected radial endpoint of the radial line based on radial endpoints of one or more additional radial lines.
18. The method of claim 15 , wherein removing the discontinuity comprises removing the radial line from the plurality of radial lines.
19. The method of claim 15 , further comprising calculating shape features of the object based on the plurality of radial endpoints using the processor.
20. A computer program product, comprising a tangible computer-readable storage medium whose contents include a program with instructions being executed on a computer system so as to perform a method for identifying a discontinuity in the boundary of an object in an image, the method comprising:
providing a system, wherein the system comprises distinct software modules, and wherein the distinct software modules comprise an image capture module and an image analysis module;
receiving a captured image of one or more objects illuminated by a light source from an imaging device using the image capture module;
calculating one or more skeletons of the one or more objects in the captured image using the image analysis module;
calculating one or more boundaries of the one or more objects in the captured image using the image analysis module;
extending a plurality of radial lines from a spine point of a skeleton of the one or more skeletons to the one or more boundaries using the image analysis module, wherein each radial line of the plurality of radial lines intersects a boundary of the one or more boundaries at a radial endpoint producing a plurality of radial endpoints;
calculating for each radial endpoint of the plurality of radial endpoints an expected radial endpoint based on two or more neighboring radial endpoints using the image analysis module; and
if the difference between the each radial endpoint and its expected radial endpoint exceeds a threshold, identifying a radial line including the each radial endpoint as a discontinuity in a boundary of an object of the one or more objects using the image analysis module.