Calibrating crop row computer vision system

Patent metadata

Field	Value
Publication number	US-11659783-B2
Application number	US-202117505302-A
Country	US
Kind code	B2
Filing date	Oct 19, 2021
Priority date	Jul 11, 2018
Publication date	May 30, 2023
Grant date	May 30, 2023

Abstract

System and techniques for calibrating a crop row computer vision system are described herein. An image set that includes crop rows and furrows is obtained. Models of the field are searched to find a model that best fits the field. A calibration parameter is extracted from the model and communicated to a receiver.

First claim

The invention claimed is:

1. A device for calibrating a crop row computer vision system, the device comprising:
a memory including instructions; and
processing circuitry that is configured by the instructions to:
obtain a set of one or more images of a field that includes crop rows and furrows, the set of one or more images captured by a sensor associated with the computer vision system;
iteratively test a set of models on an image from the set of one or more images to select a model of the field, the test based on a correspondence between a periodic component of the model and the crop rows or the furrows in the image, an iteration including:
a selection of a period for a function as a current model;
a test of the current model for correspondence between a feature of the function and the crop rows or the furrows in the image to produce a value; and
a determination of whether the value is better than a previous value from a previous iteration, iterations terminating in response to a predefined iteration limit being reached or to the value being worse than the previous value indicating a maximum extremum is reached, the model corresponding to the maximum extremum, the period selected at a next iteration to move towards the maximum extremum based on the determination of whether the value is better than the previous value from the previous iteration;

determine a scaling of crop rows or furrows in the image using the model;
determine a calibration parameter that is indicative of a configuration of the sensor relative to the field based on the scaling of crops rows or furrows; and
communicate the calibration parameter to a receiver.

2. The device of claim 1 , wherein a convolution of the model is applied across the image to test the model.

3. The device of claim 2 , wherein the convolution is horizontal.

4. The device of claim 2 , wherein the convolution is vertical.

5. The device of claim 4 , wherein the image is divided into horizontal segments with different vertical positions in the image, and wherein the model is tested on multiple of the horizontal segments to produce the convolution.

6. The device of claim 5 , wherein the processing circuitry is configured to perform the test of the model on the multiple horizontal segments in parallel.

7. The device of claim 5 , wherein the calibration parameter is a pitch of a sensor that captured the set of one or more images, and wherein the pitch is determined by different periods selected as maximum extremum in different horizontal segments.

8. The device of claim 5 , wherein a horizontal segment is a scanline of the image.

9. The device of claim 1 , wherein the function is a wave.

10. The device of claim 9 , wherein the feature of the function is a top of the wave or a bottom of the wave.

11. The device of claim 1 , wherein the calibration parameter is a height of a sensor that captured the set of one or more images.

12. The device of claim 1 , wherein the calibration parameter is used to transform measurements made on images captured from the sensor during navigation into navigational measurements, via a homography, used to control movement of agricultural equipment.

13. A method for calibrating a crop row computer vision system, the method comprising:
obtaining a set of one or more images of a field that includes crop rows and furrows, the set of one or more images captured by a sensor associated with the computer vision system;
iteratively testing a set of models on an image from the set of one or more images to select a model of the field, the testing based on a correspondence between a periodic component of the model and the crop rows or the furrows in the image, an iteration including:
selecting a period for a function as a current model;
testing the current model for correspondence between a feature of the function and the crop rows or the furrows in the image to produce a value; and
determining whether the value is better than a previous value from a previous iteration, iterations terminating in response to a predefined iteration limit being reached or to the value being worse than the previous value indicating a maximum extremum is reached, the model corresponding to the maximum extremum, the period selected at a next iteration to move towards the maximum extremum based on the determination of whether the value is better than the previous value from the previous iteration;

determining a scaling of crop rows or furrows in the image using the model;
determining a calibration parameter that is indicative of a configuration of the sensor relative to the field based on the scaling of crops rows or furrows; and
communicating the calibration parameter to a receiver.

14. The method of claim 13 , wherein a convolution of the model is applied across the image during testing.

15. The method of claim 14 , wherein the convolution is vertical.

16. The method of claim 15 wherein the image is divided into horizontal segments with different vertical positions in the image, and wherein the model is tested on multiple of the horizontal segments to produce the convolution.

17. The method of claim 16 , wherein the calibration parameter is a pitch of a sensor that captured the set of one or more images, and wherein the pitch is determined by different periods selected as maximum extremum in different horizontal segments.

18. The method of claim 13 , wherein the function is a wave, and wherein the feature of the function is a top of the wave or a bottom of the wave.

19. The method of claim 13 , wherein the calibration parameter is a height of a sensor that captured the set of one or more images.

20. The method of claim 13 , wherein the calibration parameter is used to transform measurements made on images captured from the sensor during navigation into navigational measurements, via a homography, used to control movement of agricultural equipment.

Assignees

• Raven Ind Inc

Inventors

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

Classifications

• G06V20/188

  Vegetation · CPC title

• G06T2207/10012

  Stereo images · CPC title

• G06T2207/30188

  Vegetation; Agriculture · CPC title

• G06T7/80

  Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration · CPC title

• G06T2207/30252

  Vehicle exterior; Vicinity of vehicle · CPC title