Granted Patent B2

US 12,260,346 · App. 17/143,536 · Granted Mar 25, 2025

Inventors: Angeles Casas (San Francisco, CA); Xiaoyuan Yang (Pleasanton, CA); Ho Jin Kim (San Francisco, CA); Steven Ward (Moraga, CA)

Assignee: CLIMATE LLC

G06N5/04G06N20/00

Abstract

Claims (75)

1. A computer-implemented method comprising:

receiving, at an agricultural intelligence computer system, agronomic training data, wherein the agronomic training data comprises optical remote sensing data generated by optical remote sensors for a plurality of first agronomic fields, precipitation data for the plurality of first agronomic fields, and measured field data for the plurality of first agronomic fields, the optical remote sensing data including short-wave infrared (SWIR) band data indicating reflectance as a function of wavelength;

training a machine learning model, at the agricultural intelligence computer system, using a part of the optical remote sensing data, relative water content (RWC) of crop residue in the plurality of first agronomic fields based on the precipitation data, and the measured data of the agronomic training data, wherein the machine learning model is configured to predict agronomic field property data, wherein the part of the optical remote sensing data is defined by a wavelength range of 2100-2300 nm;

in response to receiving a request from a client computing device for agronomic field property data for one or more agronomic fields, automatically predicting the agronomic field property data for the one or more agronomic fields, using the trained machine learning model, wherein the agronomic field property data comprises crop residue cover (CRC) data indicating one or more percentages of one or more ground surface residue coverages for the one or more agronomic fields;

based on the agronomic field property data, automatically generating a first graphical representation;

causing to display the first graphical representation on the client computing device.

2. The method of claim 1 , wherein the first graphical representation depicts a high erosion risk map that indicates CRC data values that are lower than a certain percentage value;

wherein the method further comprises based on agronomic field property data values of the predicted agronomic field property data, dividing the agronomic field property data into a plurality of classes, wherein each class corresponds to a range of the agronomic field property data values;

assigning an estimated tillage practice type to each class;

based on the estimated tillage practice types, generating a second graphical representation;

causing to display the second graphical representation on the client computing device;

wherein the second graphical representation depicts one of: an erosion risk map, a nutrient depletion risk map, a water runoff risk map, or a disease risk assessment map;

generating recommendations determined based on, at least in part, the high erosion risk map that indicates CRC data; wherein the recommendations comprise one or more of: tillage practice adjustments, tillage date adjustments, harvest date adjustments, plant date adjustments, crop type adjustments, risk management, or seeding recommendations;

transmitting the recommendations to a controller that controls a mechanical machine so as to cause the mechanical machine to execute the recommendations.

3. The method of claim 1 , wherein the first graphical representation comprises one or more pixel-based images;

mapping the predicted agronomic field property data onto the one or more pixel-based images, wherein each pixel of the one or more pixel-based images corresponds to an agronomic field property data value;

generating a set of colors, wherein each color corresponds to an agronomic field property data value;

assigning colors from the set of colors to each pixel based on the agronomic field property data values of each pixel.

4. The method of claim 2 , wherein the second graphical representation comprises one or more pixel-based images;

mapping the plurality of classes onto the one or more pixel-based images, wherein each pixel of the one or more pixel-based images corresponds to one or more classes of the plurality of classes;

generating a set of colors, wherein each color corresponds to one or more classes of the plurality of classes;

assigning colors from the set of colors to the pixels based on the one or more classes of the plurality of classes of each pixel.

5. The method of claim 1 , wherein the machine learning model is one of:

a Gaussian process regression model, or a multiple linear regression model.

6. The method of claim 1 , further comprising:

inputting, from the client computing device, data associated with the one or more agronomic fields, wherein the data comprises one or more of: an identifier, a crop type, field boundary definition, a location definition, or prediction time window definition.

7. The method of claim 1 , further comprising:

selecting a set of optical remote sensing data from the optical remote sensing data based on a time window;

using the set of optical remote sensing data as agronomic training data.

8. The method of claim 1 , further comprising:

based on the received optical remote sensing data, detecting contaminated optical remote sensing data from the received optical remote sensing data, wherein the contaminated optical remote sensing data are received optical remote sensing data affected by clouds;

determining whether the contaminated optical remote sensing data has reached a contamination threshold;

in response to determining that the contaminated optical remote sensing data has reached the contamination threshold, discarding the contaminated optical remote sensing data.

9. The method of claim 1 , further comprising:

receiving, at the agricultural intelligence computer system, new agronomic training data;

in response to receiving the new agronomic training data, retraining the machine learning model with the new agronomic training data.

10. One or more non-transitory computer-readable storage media storing one or more computer instructions which, when executed by one or more computer processors, cause the one or more computer processors to perform:

receiving, at an agricultural intelligence computer system, agronomic training data, wherein the agronomic training data comprises optical remote sensing data generated by optical remote sensors for a plurality of first agronomic fields, precipitation data for the plurality of first agronomic fields, and measured field data for the plurality of first agronomic fields, the optical remote sensing data including short-wave infrared (SWIR) band data indicating reflectance as a function of wavelength;

training a machine learning model, at the agricultural intelligence computer system, using a part of the optical remote sensing data, relative water content (RWC) of crop residues in the plurality of first agronomic fields based on the precipitation data, and the measured data of the agronomic training data, wherein the machine learning model id configured to predict agronomic field data, wherein the part of the optical remote sensing data is defined by a wavelength range of 2100-2300 nm;

in response to receiving a request from a client computing device for agronomic field property data for one or more agronomic fields, automatically predicting the agronomic field property data for the one or more agronomic fields, using the trained machine learning model, wherein the agronomic field property data comprises crop residue cover (CRC) data indicating one or more percentages of one or more ground surface residue coverages for the one or more agronomic fields;

based on the agronomic field property data, automatically generating a first graphical representation;

causing to display the first graphical representation on the client computing device.

11. The one or more non-transitory computer-readable storage media of claim 10 , wherein the first graphical representation depicts a high erosion risk map that indicates CRC data values that are lower than a certain percentage value;

wherein the one or more non-transitory computer-readable storage media stores additional instructions for:

based on agronomic field property data values of the predicted agronomic field property data, dividing the agronomic field property data into a plurality of classes, wherein each class corresponds to a range of the agronomic field property data values;

assigning an estimated tillage practice type to each class;

based on the estimated tillage practice types, generating a second graphical representation;

causing to display the second graphical representation on the client computing device;

wherein the second graphical representation depicts one of: an erosion risk map, a nutrient depletion risk map, a water runoff risk map, or a disease risk assessment map;

generating recommendations determined based on, at least in part, the high erosion risk map that indicates CRC data; wherein the recommendations comprise one or more of: tillage practice adjustments, tillage date adjustments, harvest date adjustments, plant date adjustments, crop type adjustments, risk management, or seeding recommendations;

transmitting the recommendations to a controller that controls a mechanical machine so as to cause the mechanical machine to execute the recommendations.

12. The one or more non-transitory computer-readable storage media of claim 10 , wherein the first graphical representation comprises one or more pixel-based images;

mapping the predicted agronomic field property data onto the one or more pixel-based images, wherein each pixel of the one or more pixel-based images corresponds to an agronomic field property data value;

generating a set of colors, wherein each color corresponds to an agronomic field property data value;

assigning colors from the set of colors to each pixel based on the agronomic field property data values of each pixel.

13. The one or more non-transitory computer-readable storage media of claim 11 , wherein the second graphical representation comprises one or more pixel-based images;

mapping the plurality of classes onto the one or more pixel-based images, wherein each pixel of the one or more pixel-based images corresponds to one or more classes of the plurality of classes;

generating a set of colors, wherein each color corresponds to one or more classes of the plurality of classes;

assigning colors from the set of colors to the pixels based on the one or more classes of the plurality of classes of each pixel.

14. The one or more non-transitory computer-readable storage media of claim 10 , wherein the machine learning model is one of:

a Gaussian process regression model, or a multiple linear regression model.

15. The one or more non-transitory computer-readable storage media of claim 10 , storing additional instructions for:

inputting, from the client computing device, data associated with the one or more agronomic fields, wherein the data comprises one or more of: identifier data, crop type data, field boundary data, location data, or prediction time window data.

16. The one or more non-transitory computer-readable storage media of claim 10 , storing additional instructions for:

selecting a set of optical remote sensing data from the optical remote sensing data based on a time window;

using the set of optical remote sensing data as agronomic training data.

17. The one or more non-transitory computer-readable storage media of claim 10 , storing additional instructions for:

based on the received optical remote sensing data, detecting contaminated optical remote sensing data from the received optical remote sensing data, wherein the contaminated optical remote sensing data are received optical remote sensing data affected by clouds;

determining whether the contaminated optical remote sensing data reaches a contamination threshold; and

discarding the contaminated optical remote sensing data where the contamination threshold is reached.

18. The one or more non-transitory computer-readable storage media of claim 10 , storing additional instructions for:

receiving, at the agricultural intelligence computer system, new agronomic training data; and

in response to receiving the new agronomic training data, retraining the machine learning model with the new agronomic training data periodically.

19. The method of claim 1 , wherein training the machine learning model further uses a part of the optical remote sensing data define by a wavelength range of 1600-1700 nm.

20. The one or more non-transitory computer-readable storage media of claim 10 , wherein training the machine learning model further uses a part of the optical remote sensing data defined by a wavelength range of 1600-1700 nm.

Assignments (6)

Continuity (2)

References Cited (19)