Image filter based on row identification

1. A system to filter an image, the system comprising:

an image capture controller to:

interface with a camera mounted on agricultural equipment; and

obtain an image of a crop row, from the camera, during operation of the agricultural equipment corresponding to a field with the crop row;

a crop row characteristic circuit set to:

calculate a crop row center from the image of the crop row; and

obtain a set of expected crop characteristics of the crop row; and

a filter circuit set to:

obtain a filter corresponding to the set of expected crop characteristics of the crop row; and

categorize elements in the image into categories based on applying the filter to the image, the filter anchored on the crop row center, including the filter circuit set to:

create a mask for the image, the mask anchored to the crop row center with edges defined by the filter;

apply the mask to the image; and

assign a category of crop element or non-crop element to elements of the image based on application of the mask to the image.

2. The system of claim 1 , wherein to determine the crop row center includes the crop row characteristic circuit set to:

apply a feature threshold to identify elements of interest in the image; and

perform a Hough transform on the elements of interest to generate a crop row center line.

3. The system of claim 1 , wherein the a set of expected crop characteristics are determined from an average of previously determined crop characteristics corresponding to a plurality of previous images of crops of a same type as the crop row.

4. The system of claim 1 , wherein the filter is transformed based on a perspective of a sensor that captured the image.

5. The system of claim 1 comprising a discriminator to:

compare non-crop elements, as determined by the filter circuit set against a desired noncrop element model to determine a deviation in the non-crop elements; and

flag the deviation.

6. The system of claim 5 comprising an applicator to control crop product application based on the flagged deviation.

7. The system of claim 6 , wherein the crop product application is contemporaneous to capturing the image.

8. A method to filter an image, the method comprising:

calculating a crop row center represented in an image of the crop row;

obtaining a filter corresponding to a set of expected crop characteristics of the crop row; and

categorizing elements in the image into categories based on applying the filter to the image, the filter anchored on the crop row center, including:

creating a mask for the image, the mask anchored to the crop row center with edges defined by the filter;

applying the mask to the image; and

assigning a category of crop element or non-crop element to elements of the image based on application of the mask to the image.

9. The method of claim 8 , wherein determining the crop row center includes:

applying a feature threshold to identify elements of interest in the image; and

performing a Hough transform on the elements of interest to generate a crop row center line.

10. The method of claim 8 , wherein the a set of expected crop characteristics are determined from an average of previously determined crop characteristics corresponding to a plurality of previous images of crops of a same type as the crop row.

11. The method of claim 8 , wherein the filter is transformed based on a perspective of a sensor that captured the image.

12. The method of claim 8 comprising:

comparing non-crop elements against a desired non-crop element model to determine a deviation in the non-crop elements; and

flagging the deviation.

13. The method of claim 12 comprising controlling crop product application based on the flagged deviation.

14. The method of claim 13 , wherein the crop product application is contemporaneous to capturing the image.

15. A set of non-transitory machine readable media including instructions that, when executed by a machine, cause the machine to perform operations comprising:

calculating a crop row center represented in an image of the crop row;

obtaining a filter corresponding to a set of expected crop characteristics of the crop row; and

categorizing elements in the image into categories based on applying the filter to the image, the filter anchored on the crop row center, including:

creating a mask for the image, the mask anchored to the crop row center with edges defined by the filter;

applying the mask to the image; and

assigning a category of crop element or non-crop element to elements of the image based on application of the mask to the image.

16. The set of machine readable media of claim 15 , wherein determining the crop row center includes:

applying a feature threshold to identify elements of interest in the image; and

performing a Hough transform on the elements of interest to generate a crop row center line.

17. The set of machine readable media of claim 15 , wherein the a set of expected crop characteristics are determined from an average of previously determined crop characteristics corresponding to a plurality of previous images of crops of a same type as the crop row.

18. The set of machine readable media of claim 15 , wherein the filter is transformed based on a perspective of a sensor that captured the image.

19. The set of machine readable media of claim 15 , wherein the operations further comprise:

comparing non-crop elements against a desired non-crop element model to determine a deviation in the non-crop elements; and

flagging the deviation.

20. The set of machine readable media of claim 19 , wherein the operations further comprise controlling crop product application based on the flagged deviation.

21. The set of machine readable media of claim 20 , wherein the crop product application is contemporaneous to capturing the image.