SYSTEM FOR VEHICULAR GUIDANCE WITH RESPECT TO HARVESTED CROP

• US 20120101695A1
• Filed: 01/04/2012
• Published: 04/26/2012
• Est. Priority Date: 07/01/2005
• Status: Active Grant

• Current Assignee: Deere & Company
• Original Assignee: Deere & Company
• Inventors: Han, Shufeng, Wei, Jiantao, Rovira-Mas, Francisco
• Granted Patent: US 8,433,483 B2
• Time in Patent Office: Days
• US Class Current: 701/50
• CPC Class Codes: A01B 69/001 Steering by means of optical assistance, e.g. television cameras steering devices for road marking vehicles E01C23/163
• CPC Class Codes: A01B 69/008 automatic

Abstract

A discriminator identifies windrow pixels associated with a windrow within a collected image. A definer defines a search space with respect to a vehicle. An evaluator determines respective spatial correlations between the defined search space and the windrow pixels for different angular displacements of the search space. An alignment detector or search engine determining a desired vehicular heading as a preferential angular displacement associated with a generally maximum spatial correlation between the defined search space and the windrow pixels. An offset calculator estimates an offset of the vehicle to a central point of the windrow or a depth axis to achieve the desired vehicle heading and desired position of the vehicle with respect to the windrow.

Claims

1. (canceled)

2. A method of guiding a vehicle, the method comprising:

identifying windrow pixels associated with a windrow within a collected color image;

identifying a primary color of the windrow pixels and one or more secondary colors of background pixels;

assigning a discrimination value to each pixel in the identified windrow pixels, wherein the discrimination value indicates whether the each pixel is crop image data or background image data, and wherein each pixel in the collected color image is associated with an intensity level;

defining a search space with respect to a vehicle, the search space containing a series of scan line segments;

determining respective spatial correlations between the defined search space and the windrow pixels for different angular displacements of the search space, wherein the determining respective spatial correlations comprises determining respective scan line intensities associated with the scan line segments for different angular displacements;

determining a desired vehicular heading as a preferential angular displacement associated with a generally maximum spatial correlation between the defined search space and the windrow pixels;

estimating a central point associated with the windrow based on pixel intensity of the windrow pixels; and

estimating an offset of the vehicle to the central point of the windrow or a depth axis to achieve the desired vehicle heading and desired position of the vehicle with respect to the windrow.

3. The method according to claim 2 wherein the determining of the respective scan line intensities comprises determining an average or mean scan line intensity for scan line segments within the search space.

4. The method according to claim 2 wherein the determining of the respective scan line intensities comprises determining a mode scan line intensity for scan line segments within the search space.

5. The method according to claim 2 wherein the determining of the respective scan line intensities comprises determining an aggregate scan line intensity for scan line segments within the search space.

6. The method according to claim 2 wherein the determining a desired vehicular heading comprises determining the desired vehicular heading as a preferential angular displacement associated with a scan line intensity within the search space meeting or exceeding a threshold scan line intensity value.

10. The method according to claim 2 further comprising:

deriving a segmented image of crop rows as the crop image data from the collected color image.

25. The method according to claim 2, further comprising:

estimating a reliability of the preferential angular heading based on compliance with a correlation value associated with the alignment of the search space and the windrow pixels, wherein the intensity level associated with one or more windrows is evaluated based on the intensity level associated with the each pixel in the collected color image.

26. The method according to claim 2, wherein the windrow pixels are identified by:

calculating the mean of the vector for each class in a number of classes, index_mean_j , wherein the number of classes include a first class representing windrow pixels and a second class representing background pixels;

calculating the covariance matrix for each class, cov_j;

calculating the average of the covariance matrix, cov_mean, for all classes; and

deriving the coefficients of discriminant function;

C_j=index_mean_j×

(cov_mean)<sup>−

1</sup>;

Co_j=−

0.5×

C_j×

index_mean_j.

27. The method according to claim 26, wherein responsive to deriving the coefficients of discriminant function, the coefficients of discriminant function are used to calculate the discriminant functions for each pixel evaluated:

DF_j=C_j×

X+Co_j,

wherein X is the vector of the each pixel evaluated.

7-9. -9. (canceled)

11-24. -24. (canceled)

28. A method of guiding a vehicle, the system comprising:

identifying windrow image data associated with a harvested crop in a collected color image and background image data in the collected color image based on color and capable of assigning a discrimination value to each portion of image data identified in the collected color image based on whether the each portion is window image data or background image data;

defining a search space with respect to a vehicle, the search space containing a series of scan line segments;

determining respective spatial correlations between the define defined_search space and the windrow image data for different angular displacements of the search space, wherein the respective spatial correlations are based on scan line intensities associated with the scan line segments for different angular displacements;

determining a desired vehicle heading as a preferential angular heading associated with a generally maximum spatial correlation between the defined search space and the windrow image data; and

identifying a desired offset of the vehicle that is generally aligned with respect to the windrow, associated with the collected color image, consistent with the desired vehicle heading.

29. The method of claim 28, wherein the each portion of image data identified is selected from one of pixels and bit maps.