Adaptive color transformation to aid computer vision

Patent metadata

Field	Value
Publication number	US-11606478-B2
Application number	US-202117512330-A
Country	US
Kind code	B2
Filing date	Oct 27, 2021
Priority date	Jul 11, 2018
Publication date	Mar 14, 2023
Grant date	Mar 14, 2023

Abstract

System and techniques for adaptive color transformation to aid computer vision are described herein. Colors from an image are mapped into a multi-dimensional space to create a distribution of colors in the image. A line can be fit to the distribution. Here, the line includes an angle relative to a coordinate system of the multi-dimensional space. A transformation to colors can then be applied to the image based on the angle of the line. The transformation producing a reduced image where a color complexity of the original image is reduced.

First claim

The invention claimed is:

1. A system for adaptive color transformation to aid computer vision, the system comprising:
a first sensor configured for coupling with an agricultural vehicle, the first sensor configured to generate a color image of at least one row of an agricultural field having one or more crop rows and intervening furrows;
a second sensor configured for coupling with the agricultural vehicle, the second sensor configured to generate a disparity image corresponding to the color image, the disparity image including distance data indicative of distances between the second sensor and the agricultural field;
a processor in communication with the first and second sensors, the processor includes:
a mapping module configured to map colors of the color image and the distance data of the disparity image to a distribution in a multi-dimensional space, the multi-dimensional space having a first dimension corresponding to one or more constituent components of the colors and a second dimension corresponding to the distance data;
a line fitting module to fit a line to the distribution in the multi-dimensional space;
a filter module to generate a reduced image having a combined image dimension, the combined image dimension obtained by using an angle between the line and the first dimensions or the second dimension to map the one or more constituent components of the colors or the distance data to the combined image dimension; and
a navigation controller in communication with the processor, the navigation controller configured to receive the reduced image and conduct guidance of the agricultural vehicle based on the reduced image and combined image dimension.

2. The system of claim 1 , wherein the first and second sensors are a same sensor configured to detect color and distance information.

3. The system of claim 1 , wherein each pixel in the disparity image is indicative of a distance between the second sensor and a corresponding region in the agricultural field.

4. The system of claim 1 , wherein using angle between the line and the first dimensions or the second dimension to map the one or more constituent components of the colors and the distance data to the combined image dimension increases a contrast between the one or more constituent components of the colors or the distance data in the reduced image.

5. The system of claim 4 , wherein the navigation controller is configured to discern a crop row from a furrow based on the increased contrast between the one or more constituent components of the colors or the distance data in the reduced image.

6. The system of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to apply a linear regression to the distribution to find the line.

7. The System of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to perform an angular scan from a centroid of the distribution to find the line.

8. The system of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to apply a theta distribution, wherein a coordinate system of the multi-dimensional space is a polar coordinate system.

9. The system of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to apply a random consensus sample (RANSAC) line model to the distribution.

10. The system of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to apply a standard deviational ellipse to the distribution.

11. The system of claim 1 , wherein, to fit the line to the distribution, the line fitting module includes a module to:
cluster points of the distribution; and
connect a first centroid of a first cluster to a second centroid of a second cluster to find the line.

12. The system of claim 1 , wherein the color image is preprocessed from a first color representation that does not include a luminance component of the colors of the color image to a second color representation that includes a luminance component.

13. A system for adaptive image transformation to aid computer vision, the system comprising:
one or more sensors configured for coupling with an agricultural vehicle, the one or more sensors configured to generate an image of at least one row of an agricultural field, the image having pixels, each of the pixels includes a positional element and an informational element, the agricultural field having one or more crop rows and intervening furrows;
a processor in communication with the one or more sensors, the processor includes:
a mapping module configured to map the informational elements of the pixels of the image to a distribution in a multi-dimensional space, the multi-dimensional space having a first dimension corresponding to a first constituent component of the informational element of the pixels and a second dimension corresponding to a second constituent component of the informational element of the pixels;
a line fitting module to fit a line to the distribution in the multi-dimensional space;
a filter module to generate a reduced image having a combined image dimension, the combined image dimension obtained by using an angle between the line and the first or second dimensions to map the first and second constituent components of the informational elements of pixels to the combined image dimension; and

a navigation controller in communication with the processor, the navigation controller configured to receive the reduced image and conduct guidance of the agricultural vehicle based on the reduced image and combined image dimension.

14. The system of claim 13 , wherein using angle between the line and the first dimensions or the second dimension to map the first and second constituent components of the informational element of pixels to the combined image dimension increases a contrast between the informational element of pixels in the reduced image.

15. The system of claim 13 , wherein the one or more sensors include a color image sensor, and the first and second constituent components of the informational element of the pixels are constituent components of colors of the pixels.

16. The system of claim 13 , wherein the one or more sensors include a depth sensor, and the first and second constituent components of the informational element of the pixels are distances to the agricultural field.

17. The system of claim 14 , wherein:
the one or more sensors include a color image sensor and a disparity sensor;
the first constituent component of the informational element of the pixels include a constituent component of colors of the pixels; and
the second constituent component of the informational element of the pixels include distances to the agricultural field.

18. The system of claim 17 , wherein the color image sensor and the disparity sensor are a same sensor configured to detect color and distance.

19. The system of claim 14 , wherein the navigation controller is configured to discern a crop row from a furrow based on the increased contrast between the informational element of pixels in the reduced image.

20. A method for adaptive color transformation to aid computer vision, the method comprising:
obtaining a color image of at least one row of an agricultural field having one or more crop rows or intervening furrows;
obtaining a disparity image corresponding to the color image, the disparity image including distance data indicative of distances between a sensor that captured the dispa

Assignees

• Raven Ind Inc

Inventors

• Sneyders Yuri
• Preheim John D
• Van Roekel Jeffrey Allen

Classifications

• A01B69/001

  Steering by means of optical assistance, e.g. television cameras (steering devices for road marking vehicles E01C23/163) · CPC title

• G06T2207/30188

  Vegetation; Agriculture · CPC title

• H04N1/6063 Primary

  dependent on the contents of the image to be reproduced · CPC title

• G06V20/56

  exterior to a vehicle by using sensors mounted on the vehicle · CPC title

• G06V10/56

  relating to colour · CPC title