Machine vision calibration with cloud computing systems

Patent metadata

Field	Value
Publication number	US-8965104-B1
Application number	US-201213601319-A
Country	US
Kind code	B1
Filing date	Aug 31, 2012
Priority date	Feb 10, 2012
Publication date	Feb 24, 2015
Grant date	Feb 24, 2015

Abstract

A cloud computing system is configured to (i) receive image and environmental data from a computing device, (ii) apply a plurality of image processing algorithms to the received image a plurality of times to generate a corresponding plurality of image processing results, where each application of an image processing algorithm to the received image is executed with a different corresponding parameter set, and (iii) based on the image processing results, select an image processing algorithm and corresponding parameter set for the computing device to use for image processing operations. The cloud computing device may also correlate the results of its analysis with the environmental data received from the computing device, and store the correlation in a machine vision knowledge base for future reference. In some embodiments, the computing device is a component of a robot.

First claim

What is claimed is:

1. A method comprising:
receiving a first image from a computing device;
applying a plurality of different image processing algorithms to the first image a plurality of times to generate a corresponding plurality of image processing results, wherein each application of an individual image processing algorithm to the first image is executed with a different corresponding parameter set comprising one or more image processing parameters;
determining a plurality of quality scores, wherein each image processing result of the plurality of image processing results has a corresponding quality score;
selecting an image processing algorithm and a corresponding parameter set, wherein the selected image processing algorithm and parameter set correspond to an image processing result having a corresponding quality score above a threshold quality score;
instructing the computing device to use the selected image processing algorithm and the selected parameter set for image processing operations;
receiving additional data associated with the first image from the computing device, wherein the additional data comprises at least one of environmental data, task data, and object data;
associating the additional data with the selected image processing algorithm and parameter set; and
storing an indication of the association between the selected image processing algorithm, parameter set, and additional data in a machine vision knowledge base.

2. The method of claim 1 , further comprising:
sending one or both of the selected image processing algorithm and parameter set to the computing device.

3. The method of claim 1 , further comprising:
based at least in part on the additional data, determining at least one of (i) the plurality of image processing algorithms to apply to the first image and (ii) one or more sets of image processing parameters for use with an individual image processing algorithm.

4. The method of claim 1 , wherein applying the plurality of different image processing algorithms to the received first image is performed substantially in parallel by a cloud computing system.

5. The method of claim 1 , wherein determining the plurality of quality scores comprises evaluating an individual image processing result based on a set of one or more quality metrics.

6. The method of claim 1 , further comprising:
receiving a second image from the computing device;
selecting one or more of an updated image processing algorithm and/or an updated parameter set based at least in part on one of (i) an analysis of the second image and (ii) a machine vision knowledge base lookup; and
instructing the computing device to use the selected updated image processing algorithm and/or updated corresponding parameter set for image processing operations.

7. A method comprising:
receiving environmental data from a robot, wherein the environmental data includes one or more attributes associated with an area in which the robot is operating;
accessing an index stored on tangible, non-transitory computer readable media, wherein the index comprises environmental data, image processing algorithms, and parameter sets for use with image processing algorithms, and wherein individual environmental data in the index has a corresponding image processing algorithm and parameter set comprising one or more image processing parameters for use with the corresponding image processing algorithm;
selecting an image processing algorithm with a parameter set from the index that corresponds to one or more attributes of the environmental data received from the robot; and
instructing the robot to use the selected image processing algorithm and parameter set to process images obtained by one or more image sensors associated with the robot.

8. The method of claim 7 , wherein the environmental data comprises at least one of (i) an amount of light, (ii) a type of light source, (iii) a direction of the light source, (iv) a geographic location, (v) a weather condition, (vi) a background pattern in an image, and (vii) a background color in an image.

9. A method comprising:
capturing a first image with an image sensor associated with a robot;
recording first environmental data associated with an area in which the robot is operating;
sending the first image and the first environmental data to a cloud computing system;
receiving instructions from the cloud computing system to use an identified image processing algorithm configured with an identified parameter set to process images obtained by the image sensor, wherein the instructions are based at least in part on the first image or the first environmental data sent to the cloud computing system; and
executing the identified image processing algorithm configured with the identified parameter set via one or more processors associated with a machine vision system associated with the robot.

10. The method of claim 9 , further comprising:
receiving at least one of the identified image processing algorithm and the identified parameter set from the cloud computing system.

11. The method of claim 9 , further comprising:
determining a plurality of corresponding quality scores for one or more image processing results generated by applying the identified image processing algorithm configured with the identified parameter set to a corresponding plurality of images captured with the image sensor; and
sending a second image and second environmental data from the robot to the cloud computing system in response to determining at least one of (i) one quality score of the plurality of quality scores falls below a threshold, (ii) a difference between two quality scores of the plurality of quality scores exceeds a threshold, and (iii) a rate of change of quality scores over a predetermined timeframe exceeds a threshold.

12. A system comprising:
one or more communication interfaces configured to receive environmental data sent from a robot, wherein the environmental data describes an environment in which the robot is operating;
a machine vision knowledge base configured to store a plurality of environmental data and a plurality of image processing algorithms with corresponding parameter sets for use with the image processing algorithms, wherein individual environmental data in the machine vision knowledge base has a corresponding image processing algorithm and parameter set; and
one or more processors configured to (i) select an image processing algorithm with a parameter set from the machine vision knowledge base based at least in part on the environmental data received from the robot, and (ii) instruct the robot to use the selected image processing algorithm and parameter set to process images obtained via one or more image sensors associated with the robot.

13. The system of claim 12 , wherein the environmental data comprises at least one of (i) an amount of light, (ii) a type of light source, (iii) a direction of the light source, (iv) a geographic location, (v) a weather condition, (vi) a background pattern in an image, and (vii) a background color in an image.

14. The system of claim 12 , wherein the one or more communications interfaces are further configured to send one or both of the selected image processing algorithm and the parameter set to the robot.

15. A computing device comprising:
an image sensor configured to record a first image;
one or more sensors configured to record environmental data associated with an area in which the computing device is located;
a communication interface configured to (i) send the first image and the environmental data to a cloud computing system, and (ii) in r

Assignees

• Hickman Ryan M
• Bruce James R
• Google Inc

Inventors

• Hickman Ryan M
• Bruce James R

Classifications

• G06T5/00

  Image enhancement or restoration · CPC title

• G06V10/96

  Management of image or video recognition tasks · CPC title

• G06T7/00 Primary

  Image analysis · CPC title