Method and system for dynamic feature detection

• US 8,411,929 B2
• Filed: 08/01/2012
• Issued: 04/02/2013
• Est. Priority Date: 04/09/2008
• Status: Expired due to Fees

• Current Assignee: Cognex Corporation
• Original Assignee: Cognex Corporation
• Inventors: Silver, William M.
• Primary Examiner(s): Geisel, Kara E
• Assistant Examiner(s): ALLI, IYABO
• Application Number:
  US13/563,857
• Publication Number:
  US 20130028474A1
• Time in Patent Office: 244 Days
• Field of Search: 382/103, 382/195, 382141-151
• US Class Current: 382/141
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06V 10/62 relating to a temporal dimension, e.g. time-based feature extraction; Pattern tracking
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching

Abstract

Disclosed are methods and systems for dynamic feature detection of physical features of objects in the field of view of a sensor. Dynamic feature detection substantially reduces the effects of accidental alignment of physical features with the pixel grid of a digital image by using the relative motion of objects or material in and/or through the field of view to capture and process a plurality of images that correspond to a plurality of alignments. Estimates of the position, weight, and other attributes of a feature are based on an analysis of the appearance of the feature as it moves in the field of view and appears at a plurality of pixel grid alignments. The resulting reliability and accuracy is superior to prior art static feature detection systems and methods.

Claims

1. A computer program product for optically measuring a plurality of object features of an object in motion relative to a one-dimensional field of view, such that the object passes through the one-dimensional field of view, the product tangibly embodied in a non-transitory computer readable medium, the computer program product comprising instructions being operable to cause a data processing apparatus to:

receive data for a sequence of images of the one-dimensional field of view;

detect static features across substantially all of the sequence of images of the one-dimensional field of view;

receive, from a memory, a plurality of dynamic features based on the plurality of object features, the plurality of dynamic features comprising at least one measurement of a physical property of the corresponding object feature;

calculate a plurality of maps, wherein at least one of the plurality of maps calculated for an image of the sequence of images is based on the static features of the image and the corresponding dynamic features of the plurality of dynamic features for the image, the map comprising

a pose that maps between at least a point on the object and at least a point in the image; and

an association that maps the corresponding dynamic features to at least some of the static features of the image so to produce associated features; and

compute at least one updated dynamic feature based on the corresponding map, whereby improving at least one measurement of a physical property of an object feature.

2. The product of claim 1, wherein the object is of substantially unknown appearance and the product further comprises instructions being operable to cause the data processing apparatus to compute and store in the memory new members of the plurality of dynamic features in response to static features evidencing the relative motion between the object and the one-dimensional field of view.

3. The product of claim 2, wherein each of the dynamic features further comprise an experience factor;

and wherein the instructions being operable to cause the data processing apparatus to;

calculate the plurality of maps further comprises increasing the influence of at least one corresponding dynamic feature on the map as the experience factor of the at least one corresponding dynamic feature increases; and

wherein the instructions being operable to cause the data processing apparatus to compute the at least one updated dynamic feature further comprises decreasing the influence of the map on the at least one dynamic feature as the experience factor of the at least one dynamic feature increases.

4. The product of claim 3, wherein the experience factor is responsive to a count of the images that were captured since the dynamic feature of the plurality of dynamic features was created.

5. The product of claim 3, wherein

each static feature comprises a weight indicating relative confidence that the static feature is responsive to an object feature;

and

the experience factor is responsive to the weights of the static features to which the dynamic feature corresponds.

6. The product of claim 2, further comprising instructions being operable to cause the data processing apparatus to

remove a dynamic feature of the plurality of dynamic features from the memory in response to the static features evidencing relative motion between the object and the field of view.

7. The product of claim 1, wherein the object is of substantially known appearance and the stored plurality of dynamic features are predetermined responsive to the known appearance.

8. The product of claim 7, wherein the computer program product further comprises instructions being operable to cause the data processing apparatus to

create the predetermined plurality of dynamic features responsive to a training object in the field of view of the received images.

9. The product of claim 8, wherein the instructions being operable to cause the data processing apparatus to create the predetermined plurality of dynamic features further comprises

receiving data of a second sequence of images of a training object in the one-dimensional field of view, the training object in relative motion with the field of view;

detecting training static features in the second sequence of images;

receiving, from a training memory, training dynamic features based on the plurality of training object features, the training dynamic features comprising at least one measurement of a physical property of a training object feature;

calculating a plurality of training maps, one map of the plurality of training maps for one image of the second sequence of images based on the training static features of the one training image and the corresponding training dynamic features of the plurality of training dynamic features for the training image, the training map comprising

a training pose that maps between points on the training object and points in the corresponding image; and

a training association that maps the corresponding training dynamic features to at least some of the training static features of the image; and

computing an update for at least one training dynamic feature based on the corresponding training map; and

selecting the predetermined plurality of dynamic features from the plurality of training dynamic features.

10. The product of claim 1, wherein the at least one updated dynamic feature comprises an updated position of the at least one dynamic feature physical property.

11. The product of claim 10, further comprising instructions being operable to cause the data processing apparatus to

calculate at least one score indicating relative confidence that an object substantially similar in appearance to the known appearance is in the field of view;

and

detect the presence or absence of an object substantially similar in appearance to the known appearance based on the at least one score.

12. The product of claim 10, wherein the updated position of at least one dynamic feature varies without regard to relative distance between the at least one dynamic feature and a next dynamic feature of the plurality of dynamic features.

13. The product of claim 1, wherein the data of the sequence of images comprises capture times of the sequence of images, and the product further comprises instructions being operable to cause the data processing apparatus to

calculate a velocity by dividing a change of position of the object between two images of the sequence of images by the difference in capture times of the two images;

calculate a coarse position of the object in a subsequent image using the velocity and the capture time of the subsequent image.

14. The product of claim 1, wherein the static features and the plurality of dynamic features comprise an edge position and an edge polarity.

15. The product of any of claims 1, 5, 7 and 11, further comprising instructions being operable to cause the data processing apparatus to

provide a value for consumption by an input/output module comprising information about the motion of the object, the motion signal responsive to at least one degree of freedom of a pose of the object.