Agricultural systems and methods using image quality metrics for vision-based detection of surface conditions
In one aspect, an agricultural method for monitoring surface conditions for an agricultural field includes receiving, with a computing system, an image of an imaged portion of an agricultural field, with the imaged portion of the agricultural field being represented by a plurality of pixels within the image. The method also includes identifying, with the computing system, at least one pixel-related parameter associated with the plurality of pixels within the image, determining, with the computing system, whether at least one image quality metric for the image is satisfied based at least in part on the at least one pixel-related parameter, and estimating, with the computing system, a surface condition associated with the agricultural field based at least in part on the image when it is determined that the at least one image quality metric is satisfied.
1. An agricultural system for monitoring surface conditions for an agricultural field, the system comprising:
an agricultural machine configured to travel across an agricultural field;
one or more imaging devices supported relative to the agricultural machine, the one or more imaging devices being configured to capture images of the agricultural field as the agricultural machine travels across the agricultural field;
a computing system communicatively coupled to the one or more imaging devices, the computing system being configured to:
receive, from the one or more imaging devices, an image of an imaged portion of the agricultural field, the imaged portion of the agricultural field being represented by a plurality of pixels within the image;
determine whether the image comprises a depth image or a two-dimensional image of the imaged portion of the agricultural field;
select at least one image quality metric to be applied for the image based on the determination of whether the image comprises the depth image or the two-dimensional image;
identify at least one pixel-related parameter associated with the plurality of pixels within the image;
determine whether the at least one image quality metric for the image is satisfied based at least in part on the at least one pixel-related parameter; and
estimate a surface condition associated with the agricultural field based at least in part on the image when it is determined that the at least one image quality metric is satisfied.
2. The agricultural system of claim 1 , wherein the computing system is further configured to disregard the image for purposes of estimating the surface condition when it is determined that the at least one image quality metric is not satisfied.
3. The agricultural system of claim 1 , wherein;
the one or more imaging devices comprise a stereo camera assembly;
the image comprises the depth image of the imaged portion of the agricultural field; and
the at least one pixel-related parameter comprises a completeness percentage determined based at least in part on a number of the plurality of pixels within the depth image that include depth information associated therewith.
4. The agricultural system of claim 3 , wherein the computing system is configured to compare the completeness percentage to a predetermined completeness percentage threshold and determine that the at least one quality metric is satisfied when the completeness percentage is equal to or exceeds the predetermined completeness percentage threshold.
5. The agricultural system of claim 1 , wherein the image comprises the two-dimensional image of the imaged portion of the agricultural field and the at least one pixel parameter comprises a coefficient of variation value for the image determined as a function of pixel values for the plurality of pixels within the image, the computing system being configured to compare the coefficient of variation value for the image to a predetermined minimum coefficient of variation threshold, and determine that the at least one image quality metric is satisfied when the coefficient of variation value for the image is greater than or equal to the predetermined minimum coefficient of variation threshold.
6. The agricultural system of claim 1 , wherein the image comprises the two-dimensional image of the imaged portion of the agricultural field and the at least one pixel parameter comprises a standard deviation value for the image determined as a function of pixel values for the plurality of pixels within the image, the computing system being configured to compare the standard deviation value for the image to a predetermined minimum standard deviation threshold, and determine that the at least one image quality metric is satisfied when the standard deviation value for the image is greater than or equal to the predetermined minimum standard deviation threshold.
7. The agricultural system of claim 1 , wherein the image comprises the two-dimensional image of the imaged portion of the agricultural field and the at least one pixel parameter comprises a saturation parameter for the image determined as a function of pixel values for the plurality of pixels within the image, the computing system being configured to compare the saturation parameter for the image to a predetermined maximum saturation threshold, and determine that the at least one image quality metric is satisfied when the saturation parameter for the image is less than the predetermined maximum saturation threshold.
8. The agricultural system of claim 7 , wherein the computing system is configured to determine both a mean pixel value and a standard deviation value for the image based on the pixel values for the plurality of pixels, the computing system being further configured to determine the saturation parameter as a function of the mean pixel value and the standard deviation value for the image.
9. The agricultural system of claim 1 , wherein:
the computing system is configured to determine a plurality of pixel-related parameters for the image, the plurality of pixel-related parameters comprising a coefficient of variation value, a standard deviation value, and a saturation parameter for the image determined as a function of pixel values for the plurality of pixels within the image;
the at least one image quality metric comprises at least one 2-D-related image quality metric; and
the computing system is configured to determine that the at least one 2-D-related image quality metric is satisfied when:
(1) the coefficient of variation value for the image is greater than or equal to a predetermined minimum coefficient of variation threshold;
(2) the standard deviation value for the image is greater than or equal to a predetermined minimum standard deviation threshold; and
(3) the saturation parameter for the image is less than a predetermined maximum saturation threshold.
10. The agricultural system of claim 9 , wherein:
the image comprises a color depth image of the imaged portion of the agricultural field;
the plurality of pixel-related parameters further comprises a completeness percentage determined based at least in part on a number of the plurality of pixels within the depth image that include depth information associated therewith;
the at least one image quality metric is selected initially such that the least one image quality metric comprises a depth-related image quality metric;
the computing system is initially configured to compare the completeness percentage to a predetermined completeness percentage threshold, and determine that the depth-related image quality metric is satisfied when the completeness percentage is greater than or equal to the predetermined completeness percentage threshold and;
in the event it is determined that the depth-related image quality metric is satisfied, the computing system is then configured to determine whether the at least one 2-D-related image quality metric is satisfied.
11. The system of claim 1 , wherein the at least one image quality metric to be selected differs depending on whether it is determined that the image comprises the depth image or the two-dimensional image.
12. An agricultural method for monitoring surface conditions for an agricultural field, the method comprising:
receiving, with a computing system, an image of an imaged portion of an agricultural field, the imaged portion of the agricultural field being represented by a plurality of pixels within the image;
determining, with the computing system, whether the image comprises a depth image or a two-dimensional image of the imaged portion of the agricultural field;
selecting, with the computing system, at least one image quality metric to be applied for the image based on the determination of whether the image comprises the depth image or the two-dimensional image;
identifying, with the computing system, at least one pixel-related parameter associated with the plurality of pixels within the image;
determining, with the computing system, whether the at least one image quality metric for the image is satisfied based at least in part on the at least one pixel-related parameter; and
estimating, with the computing system, a surface condition associated with the agricultural field based at least in part on the image when it is determined that the at least one image quality metric is satisfied.
13. The agricultural method of claim 12 , further comprising disregarding the image for purposes of estimating the surface condition when it is determined that the at least one image quality metric is not satisfied.
14. The agricultural method of claim 12 , further comprising initiating a control action to generate an operator notification when it is determined that the at least one image quality metric is not satisfied.
15. The agricultural method of claim 12 , wherein receiving the image of the imaged portion of the agricultural field comprises receiving the depth image of the imaged portion of the agricultural field and wherein identifying the at least one pixel-related parameter comprises determining a completeness percentage based at least in part on a number of the plurality of pixels within the depth image that include depth information associated therewith.
16. The agricultural method of claim 15 , wherein selecting the at least one image quality metric comprises selecting a depth-related image quality metric based on the image comprising the depth image and wherein determining whether the at least one image quality metric is satisfied comprises:
comparing the completeness percentage to a predetermined completeness percentage threshold; and
determining that the depth-related image quality metric is satisfied when the completeness percentage is greater than or equal to the predetermined completeness percentage threshold.
17. The agricultural method of claim 12 , wherein:
the image comprises the two-dimensional image of the imaged portion of the agricultural field;
identifying the at least one pixel-related parameter comprises determining a coefficient of variation value for the image as a function of pixel values for the plurality of pixels within the image;
selecting the at least one image quality metric comprises selecting a 2-D-related image quality metric based on the image comprising the two-dimensional image; and
determining whether the at least one image quality metric is satisfied comprises:
comparing the coefficient of variation value for the image to a predetermined minimum coefficient of variation threshold; and
determining that the 2-D-related image quality metric is satisfied when the coefficient of variation value for the image is greater than or equal to the predetermined minimum coefficient of variation threshold.
18. The agricultural method of claim 12 , wherein:
the image comprises the two-dimensional image of the imaged portion of the agricultural field;
identifying the at least one pixel-related parameter comprises determining a standard deviation value for the image as a function of pixel values for the plurality of pixels within the image;
selecting the at least one image quality metric comprises selecting a 2-D-related image quality metric based on the image comprising the two-dimensional image; and
determining whether the at least one image quality metric is satisfied comprises:
comparing the standard deviation value for the image to a predetermined minimum standard deviation threshold; and
determining that the 2-D-related image quality metric is satisfied when the standard deviation value for the image is greater than or equal to the predetermined minimum standard deviation threshold.
19. The agricultural method of claim 12 , wherein:
the image comprises the two-dimensional image of the imaged portion of the agricultural field;
identifying the at least one pixel-related parameter comprises determining a saturation parameter for the image as a function of pixel values for the plurality of pixels within the image;
selecting the at least one image quality metric comprises selecting a 2-D-related image quality metric based on the image comprising the two-dimensional image; and
determining whether the at least one image quality metric is satisfied comprises:
comparing the saturation parameter for the image to a predetermined maximum saturation threshold; and
determining that the 2-D-related image quality metric is satisfied when the saturation parameter for the image is less than the predetermined maximum saturation threshold.
20. The agricultural method of claim 12 , wherein:
identifying the at least one pixel-related parameter comprises determining a coefficient of variation, a standard deviation, and a saturation parameter for the image as a function of pixel values for the plurality of pixels within the image;
selecting the at least one image quality metric comprises selecting at least one 2-D-related image quality metric; and
determining whether the at least one image quality metric is satisfied comprises:
determining that the at least one 2-D-related image quality metric is satisfied when:
(1) the coefficient of variation value for the image is greater than or equal to a predetermined minimum coefficient of variation threshold;
(2) the standard deviation value for the image is greater than or equal to a predetermined minimum standard deviation threshold; and
(3) the saturation parameter for the image is less than a predetermined maximum saturation threshold.
21. The agricultural method of claim 20 , wherein:
receiving the image of the imaged portion of the agricultural field comprises receiving a color depth image of the imaged portion of the agricultural field;
identifying the at least one pixel-related parameter further comprises determining a completeness percentage based at least in part on a number of the plurality of pixels within the color depth image that include depth information associated therewith;
selecting the at least one image quality metric comprises initially selecting a depth-related image quality metric based on the image comprising the color depth image; and
determining whether the at least one image quality metric is satisfied comprises:
comparing the completeness percentage to a predetermined completeness percentage threshold;
determining that the depth-related image quality metric is satisfied when the completeness percentage is greater than or equal to predetermined completeness percentage threshold; and
in the event it is determined that the depth-related image quality metric is satisfied, determining whether the at least one 2-D-related image quality metric is satisfied.