Row guidance parameterization with Hough transform

1. A non-transitory machine-readable medium (MMRM) including instructions that, when executed by a machine, cause the machine to perform operations comprising:

receiving an electronic representation of a field (ERF) including a set of feature sets, the set of feature sets including at least one of a set of crop row features and a set of furrow features;

producing a first parameter space by performing a slope-intercept Hough transform (SLIHT) on members of a feature set in the set of feature sets, the first parameter space including a first slope parameter and a first intercept parameter;

identifying peaks in the first parameter space;

producing a second parameter space by performing the SLIHT on the peaks;

calculating a vanishing point based on a vanishing point peak in the second parameter space; and

calculating a track-angle error (TKE) from the vanishing point.

2. The MMRM of claim 1 , wherein producing the second parameter space includes:

producing a third parameter space by performing the SLIHT on members of a second feature set in the set of feature sets, the feature set and the second feature set being different members of the set of feature sets;

identifying second peaks in the third parameter space; and

adjusting the second parameter space with a result from performing SLIHT on the second peaks.

3. The MMRM of claim 1 , wherein receiving the ERF includes:

receiving a digital image of the field; and

applying a transform to the digital image to produce the ERF.

4. The MMRM of claim 3 , wherein the transform includes color modeling.

5. The MMRM of claim 3 , wherein the transform includes downsampling.

6. The MMRM of claim 5 , wherein the downsampling includes image distortion correction.

7. The MMRM of claim 3 , wherein the transform includes removing members of the feature set.

8. The MMRM of claim 1 , wherein calculating the TKE from the vanishing point includes:

calculating a view dip angle, the view dip angle being a vertical declination of an image capture device from horizontal when an image of the field was taken, the ERF being derived from the image of the field; and

modifying an ERF-track-angle error using the view dip angle to adjust for distortions between elements of the image and corresponding elements of the field.

9. The MMRM of claim 1 , wherein the instructions, when executed, cause the machine to perform operations comprising:

identifying a subset of peaks in the first parameter space based on a line in the first parameter space defined by the vanishing point peak;

calculating a set of intermediate intercept differences from member pairs in the subset of peaks, the intercept differences calculated on an XTK line;

calculating a single intercept difference in the ERF based on the set of intermediate intercept differences; and

calculating a cross-track distance (XTK) based on the single intercept difference.

10. The MMRM of claim 9 , wherein calculating the XTK includes applying a set of scaling operations.

11. The MMRM of claim 9 , wherein calculating the XTK includes:

calculating a camera intercept on the XTK line for a line passing through both the vanishing point and a camera position; and

calculating a set of differences, each member of the set of differences being a difference between the camera intercept and an intercept for a line represented by a member of the subset of peaks in the first parameter space;

applying a modulus operation to each member of the set of differences using the single intercept difference as a parameter to produce a set of results; and

combining the members of the set of results.

12. A system comprising:

a scene module arranged to receive an electronic representation of a field (ERF) including a set of feature sets, the set of feature sets including at least one of a set of crop row features and a set of furrow features;

a transform module arranged to:

produce a first parameter space by performing a slope-intercept Hough transform (SLIHT) on members of a feature set in the set of feature sets, the first parameter space including a first slope parameter and a first intercept parameter;

identify peaks in the first parameter space; and

produce a second parameter space by performing the SLIHT on the peaks; and

a track-angle error module arranged to:

calculate a vanishing point based on a vanishing point peak in the second parameter space; and

calculate a track-angle error (TKE) from the vanishing point.

13. The system of claim 12 , wherein to produce the second parameter space includes the transform module arranged to:

produce a third parameter space by performing the SLIHT on members of a second feature set in the set of feature sets, the feature set and the second feature set being different members of the set of feature sets;

identify second peaks in the third parameter space; and

adjust the second parameter space with a result from performing SLIHT on the second peaks.

14. The system of claim 12 , wherein to receive the ERF includes the scene module arranged to:

receive a digital image of the field; and

apply a transform to the digital image to produce the ERF.

15. The system of claim 14 , wherein the transform includes color modeling.

16. The system of claim 14 , wherein the transform includes downsampling.

17. The system of claim 16 , wherein the downsampling includes image distortion correction.

18. The system of claim 14 , wherein the transform includes removing members of a feature set.

19. The system of claim 12 , wherein to calculate the TKE from the vanishing point includes the track-angle error module arranged to:

calculate a view dip angle, the view dip angle being a vertical declination of an image capture device from horizontal when an image of the field was taken, the ERF being derived from the image of the field; and

modify an ERF-track-angle error using the view dip angle to adjust for distortions between elements of the image and corresponding elements of the field.

20. The system of claim 12 comprising a cross-track distance module arranged to:

identify a subset of peaks in the first parameter space based on a line in the first parameter space defined by the vanishing point peak;

calculate a set of intermediate intercept differences from member pairs in the subset of peaks, the intercept differences calculated on an XTK line;

calculate a single intercept difference in the ERF based on the set of intermediate intercept differences; and

calculate a cross-track distance (XTK) based on the single intercept difference.

21. The system of claim 20 , wherein to calculate the XTK includes the cross-track distance module arranged to apply a set of scaling operations.

22. The system of claim 20 , wherein to calculate the XTK includes the cross-track distance module arranged to:

calculate a camera intercept on the XTK line for a line passing through both the vanishing point and a camera position;

calculate a set of differences, each member of the set of differences being a difference between the camera intercept and an intercept for a line represented by a member of the subset of peaks in the first parameter space;

apply a modulus operation to each member of the set of differences using the single intercept difference as a parameter to produce a set of results; and

combine members of the set of results.

23. A method comprising:

receiving an electronic representation of a field (ERF) including a set of feature sets, the set of feature sets including at least one of a set of crop row features and a set of furrow features;

producing a first parameter space by performing a slope-intercept Hough transform (SLIHT) on members of a feature set in the set of feature sets, the first parameter space including a first slope parameter and a first intercept parameter;

identifying peaks in the first parameter space;

producing a second parameter space by performing the SLIHT on the peaks;

calculating a vanishing point based on a vanishing point peak in the second parameter space; and

calculating a track-angle error (TKE) from the vanishing point.

24. The method of claim 23 , wherein producing the second parameter space includes:

producing a third parameter space by performing the SLIHT on members of a second feature set in the set of feature sets, the feature set and the second feature set being different members of the set of feature sets;

identifying second peaks in the third parameter space; and

adjusting the second parameter space with a result from performing SLIHT on the second peaks.

25. The method of claim 23 , wherein receiving the ERF includes:

receiving a digital image of the field; and

applying a transform to the digital image to produce the ERF.

26. The method of claim 25 , wherein the transform includes color modeling.

27. The method of claim 25 , wherein the transform includes downsampling.

28. The method of claim 27 , wherein the downsampling includes image distortion correction.

29. The method of claim 25 , wherein the transform includes removing members of the feature set.

30. The method of claim 23 , wherein calculating the TKE from the vanishing point includes:

calculating a view dip angle, the view dip angle being a vertical declination of an image capture device from horizontal when an image of the field was taken, the ERF being derived from the image of the field; and

modifying an ERF-track-angle error using the view dip angle to adjust for distortions between elements of the image and corresponding elements of the field.

31. The method of claim 23 comprising:

identifying a subset of peaks in the first parameter space based on a line in the first parameter space defined by the vanishing point peak;

calculating a set of intermediate intercept differences from member pairs in the subset of peaks, the intercept differences calculated on an XTK line;

calculating a single intercept difference in the ERF based on the set of intermediate intercept differences; and

calculating a cross-track distance (XTK) based on the single intercept difference.

32. The method of claim 31 , wherein calculating the XTK includes applying a set of scaling operations.

33. The method of claim 31 , wherein calculating the XTK includes:

calculating a camera intercept on the XTK line for a line passing through both the vanishing point and a camera position;

calculating a set of differences, each member of the set of differences being a difference between the camera intercept and an intercept for a line represented by a member of the subset of peaks in the first parameter space;

applying a modulus operation to each member of the set of differences using the single intercept difference as a parameter to produce a set of results; and

combining the members of the set of results.