Probe placement for image processing

Patent metadata

Field	Value
Publication number	US-9995573-B2
Application number	US-201514603969-A
Country	US
Kind code	B2
Filing date	Jan 23, 2015
Priority date	Jan 23, 2015
Publication date	Jun 12, 2018
Grant date	Jun 12, 2018

Abstract

The present application discloses a probe placement module for placing probes on a virtual object depicted in an image. The probe placement module is configured to place probes on interest points of an image so that the probes can accurately represent a pattern depicted in the image. The probe placement module can be configured to place the probes so that the probes can extract balanced information on all degrees of freedom associated with the pattern's movement, which improves the accuracy of the model generated from the probes.

First claim

The invention claimed is:

1. A machine vision system comprising:
a processor configured to run a computer program stored in memory that is configured to:
determine, using an interest point detection module, a plurality of interest points from an image;
determine, by a feature distribution computation (FDC) module stored in the memory in communication with the interest point detection module, an interest point distribution of the plurality of interest points, comprising data indicative of a distribution of the interest points in the image;
determine, using a target distribution computation (TDC) module, a target distribution of a plurality of probes, wherein:
each probe from the plurality of probes is used to model a portion of a pattern; and
the target distribution is indicative of a desired placement of the probes on one or more of the interest points; and

determine, by a probe distribution match (PDM) module in communication with the TDC module and the FDC module, placements of the plurality of probes on one or more of the interest points based at least in part on the interest point distribution to match a distribution of the plurality of probes to the target distribution to generate a model for the pattern, the model comprising the plurality of probes and associated placements.

2. The machine vision system of claim 1 , wherein the processor is further configured to:
determine orientations at the plurality of interest points, and
determine a histogram of orientations indicative of a number of interest points having a particular orientation.

3. The machine vision system of claim 2 , wherein the processor is further configured to smooth the histogram of the orientations using a smoothing function.

4. The machine vision system of claim 3 , wherein the smoothing function comprises a circular kernel in an orientation domain.

5. The machine vision system of claim 4 , wherein the circular kernel comprises a cosine function.

6. The machine vision system of claim 1 , wherein the processor is further configured to find the plurality of probes that are equally balanced in all dimensions, such that each number of probes from the plurality of probes that are associated with different orientations are balanced.

7. The machine vision system of claim 1 , wherein the interest points are associated with a boundary of a pattern modeled by the plurality of probes, wherein the interest points are associated with an arc along the boundary, and wherein the processor is further configured to:
place a first probe at a first interest point;
adjust a length of an arc of the first interest point based on the number of interest points having a substantially similar orientation as the first interest point and a second number of interest points having a different orientation than the first interest point; and
place a second probe at a second interest point based on the adjusted length of the arc of the first interest point.

8. The machine vision system of claim 1 , wherein the processor is further configured to fit the distribution of the orientations of the plurality of interest points to a mixture of probability distributions comprising a mixture model.

9. The machine vision system of claim 8 , wherein the processor is further configured to determine a number of components of the mixture model of the interest point distribution by clustering the plurality of interest points into at least one cluster.

10. The machine vision system of claim 8 , wherein the processor is further configured to adjust the weights of the interest point distribution mixture model to determine the target distribution having a property that at least one of its moments is approximately equal across each of the considered dimensions.

11. The machine vision system of claim 1 , wherein the processor is further configured to use the target distribution to interest point orientation distribution ratio to re-sample the interest points and produce the plurality of probes.

12. The machine vision system of claim 1 , wherein the interest point distribution and the target distribution comprise distributions over a predetermined set of dimensions, wherein the predetermined set of dimensions comprises one of Euclidean dimensions or cylindrical dimensions.

13. A method comprising:
determining, by an interest point detection module in a machine vision system, a plurality of interest points from an image;
determining, by a feature distribution computation (FDC) module coupled to the interest point detection module in the machine vision system, an interest point distribution of the plurality of interest points, comprising data indicative of a distribution of the interest points in the image;
determining, using a target distribution computation (TDC) module, a target distribution of a plurality of probes, wherein:
each probe from the plurality of probes is used to model a portion of a pattern; and
the target distribution is indicative of a desired placement of the probes on one or more of the interest points; and

determining, by a probe distribution match (PDM) module coupled to the TDC module and the FDC module, placements of the plurality of probes on one or more of the interest points based at least in part on the interest point distribution to match a distribution of the plurality of probes to the target distribution to generate a model for the pattern, the model comprising the plurality of probes and associated placements.

14. The method of claim 13 , wherein determining the interest point distribution comprises:
determining orientations at the plurality of interest points, and
determining a histogram of the orientations indicative of a number of interest points having a particular orientation.

15. The method of claim 14 , further comprising smoothing the histogram of the orientations using a smoothing function.

16. The method of claim 13 , further comprising identifying the plurality of probes that are equally balanced in all considered dimensions, such that each number of probes from the plurality of probes that are associated with different orientations are balanced.

17. The method of claim 13 , wherein the interest points are associated with a boundary of a pattern modeled by the plurality of probes, wherein the interest points are associated with an arc along the boundary, and wherein determining the placements of the plurality of probes based at least in part on the interest point distribution comprises:
placing a first probe at a first interest point;
adjusting a length of an arc of the first interest point based on the number of interest points having a substantially similar orientation as the first interest point and a second number of interest points having a different orientation than the first interest point; and
placing a second probe at a second interest point based on the adjusted length of the arc of the first interest point.

18. The method of claim 13 , wherein determining the interest point distribution comprises fitting the plurality of interest points to a probability distribution comprising a mixture model, wherein the probability distribution is indicative of at least a distribution of orientations measured at the plurality of interest points.

19. The method of claim 13 , further comprising determining a number of components of the mixture model of the interest point distribution by clustering the plurality of interest points into at least one cluster.

20. A non-transitory computer readable medium having executable instructions associat

Assignees

• Cognex Corp

Inventors

• Barker Simon
• Michael David J
• Silver William M

Classifications

• G01B21/00 Primary

  Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant · CPC title

• G06V10/255 Primary

  Detecting or recognising potential candidate objects based on visual cues, e.g. shapes · CPC title

• G06V10/757

  Matching configurations of points or features · CPC title

• G06K9/685

  Physics · mapped topic

• G06T7/73

  using feature-based methods · CPC title