Machine-learned obstruction detection in a farming machine

1. A method for achieving a farming objective, the method comprising:

accessing, an obstruction model configured to identify obstructions in a field from image data of the field, the obstruction model generated by:

accessing image data of obstructions in a training field, each obstruction in the training field corresponding to a prescribed action occurring at a prescribed time;

labelling the image data of the obstructions in the training field, each obstruction in the image data of the obstructions in the training field labelled based on the prescribed time that the corresponding prescribed action occurred; and

training the obstruction model to identify the obstructions in the field using the labelled image data;

capturing the image data of the field including an obstruction in the field;

inputting the image data of the field into the obstruction model to identify the obstruction in the field; and

responsive to identifying the obstruction in the field, modifying treatment instructions executed by a farming machine for implementing the farming objective, execution of the modified treatment instructions causing the farming machine to avoid the obstruction in the field.

2. The method of claim 1 , further comprising:

sending a message to a mobile device of an actor, the message indicating a first prescribed action for the actor to perform in a field;

capturing image data of the actor performing the first prescribed action;

labelling the image data of the actor with the message; and

training the obstruction model on the labeled image data of the actor.

3. The method of claim 1 , wherein the prescribed actions simulate obstructions that may occur in the field and include one or more of crouching, lying down, walking, running, making a hand gesture, using an object, looking in a direction, moving an object, or wearing an item.

4. The method of claim 1 , wherein the obstruction model is a machine learning model and identifying the obstruction in the field comprises:

receiving input image data as input to the obstruction model; and

outputting a confidence score indicating a likelihood that an obstruction is present in the input image data.

5. The method of claim 1 , wherein the obstruction model comprises one or more machine learning models and inputting the image data of the field into the obstruction model further comprises:

inputting the image data of the field into the one or more machine learning models to identify obstructions; and

identifying the obstruction in the field based on outputs from the one or more machine learning models.

6. The method of claim 5 , wherein the one or more machine learning models are one or more of a pixel model, point cloud model, or a bounding box model.

7. The method of claim 1 , further comprising:

capturing image data of obstructions in the field where the farming machine implements farming objectives in the field; and

wherein the accessed image data is the captured image data.

8. The method of claim 1 , wherein the modified treatment instructions cause the farming machine to stop moving, send a notification to an external operator to get instructions based on the obstruction in the field, and proceed according to instructions from the external operator.

9. A non-transitory computer-readable storage medium storing instructions that when executed by a processor cause the processor to perform a method comprising:

accessing, an obstruction model configured to identify obstructions in a field from image data of the field, the obstruction model generated by:

accessing image data of obstructions in a training field, each obstruction in the training field corresponding to a prescribed action occurring at a prescribed time;

labelling the image data of the obstructions in the training field, each obstruction in the image data of the obstructions in the training field labelled based on the prescribed time that the corresponding prescribed action occurred; and

training the obstruction model to identify the obstructions in the field using the labelled image data;

capturing the image data of the field including an obstruction in the field;

inputting the image data of the field into the obstruction model to identify the obstruction in the field; and

responsive to identifying the obstruction in the field, modifying treatment instructions executed by a farming machine for implementing a farming objective, execution of the modified treatment instructions causing the farming machine to avoid the obstruction in the field.

10. The non-transitory computer-readable storage medium of claim 9 , wherein the instructions when executed by the processor further cause the processor to perform the method comprising:

sending a message to a mobile device of an actor, the message indicating a first prescribed action for the actor to perform in a field;

capturing image data of the actor performing the first prescribed action;

labelling the image data of the actor with the message; and

training the obstruction model on the labeled image data of the actor.

11. The non-transitory computer-readable storage medium of claim 9 , wherein the prescribed actions simulate obstructions that may occur in the field and include one or more of crouching, lying down, walking, running, making a hand gesture, using an object, looking in a direction, moving an object, or wearing an item.

12. The non-transitory computer-readable storage medium of claim 9 , wherein the obstruction model is a machine learning model and identifying the obstruction in the field comprises:

receiving input image data as input to the obstruction model; and

outputting a confidence score indicating a likelihood that an obstruction is present in the input image data.

13. The non-transitory computer-readable storage medium of claim 9 , wherein the obstruction model comprises one or more machine learning models and inputting the image data of the field into the obstruction model comprises:

inputting the image data of the field into the one or more machine learning models to identify obstructions; and

identifying the obstruction in the field based on outputs from the one or more machine learning models.

14. The non-transitory computer-readable storage medium of claim 13 , wherein the one or more machine learning models are one or more of a pixel model, point cloud model, or a bounding box model.

15. The non-transitory computer-readable storage medium of claim 9 , wherein the instructions when executed by the processor further cause the processor to perform the method comprising:

capturing image data of obstructions in the field where the farming machine implements farming objectives in the field; and

wherein the accessed image data is the captured image data.

16. The non-transitory computer-readable storage medium of claim 9 , wherein the modified treatment instructions cause the farming machine to stop moving, send a notification to an external operator to get instructions based on the obstruction in the field, and proceed according to instructions from the external operator.

17. A computer system comprising:

a processor; and

a non-transitory computer-readable storage medium storing instructions that when executed cause the processor to perform actions comprising:

accessing, an obstruction model configured to identify obstructions in a field from image data of the field, the obstruction model generated by:

accessing image data of obstructions in a training field, wherein each obstruction in the training field corresponds to a prescribed action occurring at a prescribed time;

labelling the image data of the obstructions in the training field, each obstruction in the image data of the obstructions in the training field labelled based on the prescribed time that the corresponding prescribed action occurred; and

training the obstruction model to identify the obstructions in the field using the labelled image data;

capturing the image data of the field including an obstruction in the field;

inputting the image data of the field into the obstruction model to identify the obstruction in the field; and

responsive to identifying the obstruction in the field, modifying treatment instructions executed by a farming machine for implementing a farming objective, execution of the modified treatment instructions causing the farming machine to avoid the obstruction in the field.

18. The computer system of claim 17 , the actions further comprising:

sending a message to a mobile device of an actor, the message indicating a first prescribed action for the actor to perform in a field;

capturing image data of the actor performing the first prescribed action;

labelling the image data of the actor with the message; and

training the obstruction model on the labeled image data of the actor.

19. The computer system of claim 17 , wherein the prescribed actions simulate obstructions that may occur in the field and include one or more of crouching, lying down, walking, running, making a hand gesture, using an object, looking in a direction, moving an object, or wearing an item.

20. The computer system of claim 17 , wherein the obstruction model is a machine learning model and identifying the obstruction in the field comprises:

receiving input image data as input to the obstruction model; and

outputting a confidence score indicating a likelihood that an obstruction is present in the input image data.