Methods and arrangements for configuring industrial inspection systems

Patent metadata

Field	Value
Publication number	US-10593007-B1
Application number	US-201715816098-A
Country	US
Kind code	B1
Filing date	Nov 17, 2017
Priority date	Jun 11, 2015
Publication date	Mar 17, 2020
Grant date	Mar 17, 2020

Abstract

In computer vision systems that need to decode machine-readable indicia from captured imagery, it is critical to select imaging parameters (e.g., exposure interval, exposure aperture, camera gain, intensity and duration of supplemental illumination) that best allow detection of subtle features from imagery. In illustrative embodiments, a Shannon entropy metric or a KL divergence metric is used to guide selection of an optimal set of imaging parameters. In accordance with other aspects of the technology, different strategies identify which spatial locations within captured imagery should be successively examined for machine readable indicia, in order to have a greatest likelihood of success, within a smallest interval of time. A great variety of other features and arrangements are also detailed.

First claim

The invention claimed is:

1. A camera configuration method comprising the acts:
capturing a first image of a subject with a camera, the camera employing a first combination of values for first and second imaging parameters, at least one of said first and second imaging parameters being selected from the group consisting of: exposure interval, camera gain, illumination intensity, and flash illumination interval;
computing a probability distribution measure, said probability distribution measure comprising a difference, or similarity, between a probability distribution of pixel values in the first image, and a second probability distribution;
capturing additional images of the subject with the camera, employing other combinations of values for said first and second imaging parameters, and, for each additional image, computing said probability distribution measure comprising a difference or similarity between a probability distribution of pixel values in the further image, and the second probability distribution;
identifying which combination of values for the first and second imaging parameters yields a most extreme value of said probability distribution measure; and
setting the camera to use said combination of values for the first and second imaging parameters in capturing subsequent images;
wherein said act of identifying a combination of values for the first and second imaging parameters yielding a most extreme value of said probability distribution measure has the practical effect of establishing an optimal set of imaging parameters to best allow machine detection of subtle features from imagery.

2. The method of claim 1 in which the probability distribution measure comprises an entropy metric, and the method includes identifying which combination of values for the first and second imaging parameters yielded a maximum value of said metric.

3. The method of claim 1 in which the probability distribution measure comprises a divergence metric, and the method includes identifying which combination of values for the first and second imaging parameters yielded a minimum value of said metric.

4. The method of claim 1 in which the second probability distribution is an idealized probability distribution.

5. The method of claim 1 that includes:
generating histogram data that counts pixels of different values in the first image;
determining an entropy or divergence metric from the histogram data;
for each of said additional images, generating histogram data, and determining an entropy or divergence metric from the histogram data;
identifying the combination of values for said first and second imaging parameters that yields the captured image having histogram data with the largest entropy metric, or with the smallest divergence metric.

6. The method of claim 1 that includes determining a Shannon entropy metric having the form:

H
=

∑

i
=
0

n

⁢

-

(

L
i

/
K

)

*

⁢
log
⁢

⁢
2
⁢

(

L
i

/
K

)

where n is the number of possible pixel values, K is the number of pixels in the image, and Li is the count of pixels in the image having value i.

7. The method of claim 1 in which both the first and second imaging parameters are selected from the list consisting of: exposure interval, lens aperture, camera gain, illumination intensity, and flash illumination interval.

8. The method of claim 1 in which one of said parameters comprises exposure interval, and the method includes initially determining said probability distribution measure with a first value for said exposure interval, and then decreasing the exposure interval for successive computations of said additional probability distribution measures, until a difference in said probability distribution measure between successive values of exposure interval changes sign.

9. The method of claim 1 in which said parameters comprise exposure interval and camera gain, and the method includes initially determining an entropy metric with a first value for said exposure interval and a first value for said camera gain, and thereafter alternately decreasing the exposure interval and camera gain for successive determinations of additional entropy metrics, until decreasing a first of said parameters yields a decline in entropy metric, and thereafter continuing to reduce the second of said parameters until decreasing the second parameter also yields a decline in entropy metric.

10. The method of claim 1 that includes filtering the captured images before computing said probability distribution measures, said filtering comprising, for each of plural pixels in each captured image, determining whether a value of said pixel is less than, equal to, or greater than, a value of each of eight neighboring pixels, and changing the value of said pixel in accordance with said eight determinations.

11. The method of claim 1 that includes computing said probability distribution measures for each of plural blocks in plural of said captured images, and for each block, identifying which combination of values for the first and second imaging parameters yielded the most extreme value of said probability distribution measure.

12. The method of claim 1 that further includes extracting information from one of said subsequent images, by optical character recognition, barcode decoding, image fingerprint recognition, recognition by a neural network, or watermark detection.

13. An inspection system for a production line that packages food or that applies labels to food containers, comprising:
a camera directed towards the production line to capture imagery of objects conveyed by the production line, the camera employing first and second imaging parameters, at least one of said first and second imaging parameters being selected from the group consisting of: exposure interval, camera gain, illumination intensity, and flash illumination interval; and
means for determining optimal values of said first and second imaging parameters, to best allow decoding of payload data from 2D indicia on said objects depicted in imagery captured by said camera, wherein said means includes:
software executing on a processor that causes the system to repeatedly collect image data from one or more objects, using different combinations of values for the first and second imaging parameters;
software executing on a processor that causes the system to determine, from the collected image data, for each different comb

Assignees

• Digimarc Corp

Inventors

• Long Scott M
• Haverkate Christopher M

Classifications

• H04N19/44

  Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder · CPC title

• G06T5/40

  using histogram techniques · CPC title

• H04N19/91

  Entropy coding, e.g. variable length coding [VLC] or arithmetic coding · CPC title

• G06T2201/0052

  Embedding of the watermark in the frequency domain · CPC title

• G06T1/0028 Primary

  Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking · CPC title