IP Library Granted Patent US 12,657,691
Granted Patent B2
US 12,657,691 · App. 18/344,681 · Granted Jun 16, 2026

Plant treatment model training based on agricultural image interaction

Inventors: Swen Ulrich Conrad (Mountain View, CA); Anthony John Latham (Ankeny, IA)
Assignee: Deere & Company
G06T7/001G06T2207/20081G06T2207/30188
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,657,691
App. No.
18/344,681
Granted
Jun 16, 2026
Kind
B2
Abstract

Embodiments relate to training a plant treatment model. A control system may provide images of plants for display to a user. The control system may receive one or more plant treatment action preferences of the user. The one or more plant treatment action preferences include labels for the images that identify plants in the images and treatment actions to be applied to the identified plants. The control system may train the plant treatment model based on the one or more plant treatment action preferences. The trained plant treatment model is configured to, when applied to an image of plants in a field, determine treatment actions to be applied to the plants in accordance with the one or more plant treatment action preferences of the user. The control system may configure a farming machine to operate based on the trained plant treatment model.

Claims (40)

1 . A method for training a plant treatment model, the method comprising:

providing images of plants of different sizes and types for display to and annotation by a user;

receiving one or more plant treatment action preferences of the user for a farming machine configured to perform treatment actions on plants growing in a field, the one or more plant treatment action preferences including labels for the images that identify plants in the images and treatment actions to be applied to the identified plants;

training the plant treatment model based on the one or more plant treatment action preferences of the user, the trained plant treatment model configured to, when applied to first image of one or more plants in the field, determine treatment actions to be applied to the one or more plants in accordance with the one or more plant treatment action preferences of the user,

wherein the trained plant treatment model is further configured to, when applied to a second image of multiple plants, identify a subset of the multiple plants in the second image in accordance with the one or more plant treatment action preferences of the user; and

configuring the farming machine to operate based on the trained plant treatment model.

2 . The method of claim 1 , wherein a plant identification sensitivity level of the trained plant treatment model is based on the one or more plant treatment action preferences of the user.

3 . The method of claim 1 , wherein to identify the subset of the multiple plants in the second image, the trained plant treatment model is further configured to identify weed plants with sizes in a size range in the second image and ignoring weed plants with sizes outside the size range.

4 . A method for training a plant treatment model, the method comprising:

providing images of plants of different sizes and types for display to and annotation by a user;

receiving one or more plant treatment action preferences of the user for a farming machine configured to perform treatment actions on plants growing in a field, the one or more plant treatment action preferences including labels for the images that identify plants in the images and treatment actions to be applied to the identified plants;

training the plant treatment model based on the one or more plant treatment action preferences of the user, the trained plant treatment model configured to, when applied to an image of one or more plants in the field, determine treatment actions to be applied to the one or more plants in accordance with the one or more plant treatment action preferences of the user;

configuring the farming machine to operate based on the trained plant treatment model;

receiving one or more revised plant treatment action preferences of the user; and

adjusting parameters of the trained plant treatment model based on the one or more revised plant treatment action preferences.

5 . The method of claim 4 , wherein the one or more revised plant treatment action preferences of the user are received subsequent to the user receiving indications of plant treatment actions performed by the farming machine while operating based on the trained plant treatment model.

6 . The method of claim 4 , wherein adjusting the parameters of the trained plant treatment model is performed while the farming machine is operating in the field.

7 . The method of claim 1 , wherein the labels for the images are generated based on annotations by the user.

8 . The method of claim 1 , wherein the plant treatment model includes parameters with default values, and training the plant treatment model comprises changing values of the parameters based on the labels for the images.

9 . The method of claim 1 , wherein the one or more plant treatment action preferences are generated responsive to the user annotating the provided images.

10 . A non-transitory computer-readable storage medium comprising stored instructions that, when executed by a computing device, cause the computing device to perform operations including:

providing images of plants of different sizes and types for display to and annotation by a user;

receiving one or more plant treatment action preferences of the user for a farming machine configured to perform treatment actions on plants growing in a field, the one or more plant treatment action preferences including labels for the images that identify plants in the images and treatment actions to be applied to the identified plants;

training a plant treatment model based on the one or more plant treatment action preferences of the user, the trained plant treatment model configured to, when applied to first image of one or more plants in the field, determine treatment actions to be applied to the one or more plants in accordance with the one or more plant treatment action preferences of the user,

wherein the trained plant treatment model is further configured to, when applied to a second image of multiple plants, identify a subset of the multiple plants in the second image in accordance with the one or more plant treatment action preferences of the user; and

configuring the farming machine to operate based on the trained plant treatment model.

11 . The non-transitory computer-readable storage medium of claim 10 , wherein a plant identification sensitivity level of the trained plant treatment model is based on the one or more plant treatment action preferences of the user.

12 . The non-transitory computer-readable storage medium of claim 10 , wherein to identify the subset of the multiple plants in the second image, the trained plant treatment model is further configured to identify weed plants with sizes in a size range in the second image and ignoring weed plants with sizes outside the size range.

13 . A non-transitory computer-readable storage medium comprising stored instructions that, when executed by a computing device, cause the computing device to perform operations including:

providing images of plants of different sizes and types for display to and annotation by a user;

receiving one or more plant treatment action preferences of the user for a farming machine configured to perform treatment actions on plants growing in a field, the one or more plant treatment action preferences including labels for the images that identify plants in the images and treatment actions to be applied to the identified plants;

training a plant treatment model based on the one or more plant treatment action preferences of the user, the trained plant treatment model configured to, when applied to an image of one or more plants in the field, determine treatment actions to be applied to the one or more plants in accordance with the one or more plant treatment action preferences of the user;

configuring the farming machine to operate based on the trained plant treatment model;

receiving one or more revised plant treatment action preferences of the user; and

adjusting parameters of the trained plant treatment model based on the one or more revised plant treatment action preferences.

14 . The non-transitory computer-readable storage medium of claim 13 , wherein the one or more revised plant treatment action preferences of the user are received subsequent to the user receiving indications of plant treatment actions performed by the farming machine while operating based on the trained plant treatment model.

15 . The non-transitory computer-readable storage medium of claim 13 , wherein adjusting the parameters of the trained plant treatment model is performed while the farming machine is operating in the field.

16 . The non-transitory computer-readable storage medium of claim 10 , wherein the labels for the images are generated based on annotations by the user.

17 . The non-transitory computer-readable storage medium of claim 10 , wherein the plant treatment model includes parameters with default values, and training the plant treatment model comprises changing values of the parameters based on the labels for the images.

18 . The non-transitory computer-readable storage medium of claim 10 , wherein the one or more plant treatment action preferences are generated responsive to the user annotating the provided images.

Assignments (3)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 11, 2024
From: BLUE RIVER TECHNOLOGY INC.
To: DEERE & COMPANY
Reel/Frame 069164/0195 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 31, 2023
From: CONRAD, SWEN ULRICH
To: BLUE RIVER TECHNOLOGY INC.
Reel/Frame 064441/0932 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 31, 2023
From: LATHAM, ANTHONY JOHN
To: DEERE & COMPANY
Reel/Frame 064441/0937 →
Continuity (1)
Related Publication 20250005738A1 · Jan 2, 2025
References Cited (19)
US 11580718B2 · Padwick et al. · 2023 [cited by applicant]
US 20170161560A1 · Itzhaky et al. · 2017 [cited by applicant]
US 20180330166A1 · Redden · 2018 [cited by examiner]
US 20200184128A1 · Hu · 2020 [cited by examiner]
US 20210056338A1 · Padwick et al. · 2021 [cited by applicant]
US 20210153500A1 · Kuenzi · 2021 [cited by examiner]
US 20210224927A1 · Perry et al. · 2021 [cited by applicant]
US 20220092705A1 · Khait · 2022 [cited by examiner]
US 20220183208A1 · Sibley et al. · 2022 [cited by applicant]
US 20220338421A1 · Call · 2022 [cited by examiner]
US 20230306795A1 · Lechner et al. · 2023 [cited by applicant]
US 20240172600A1 · Xu et al. · 2024 [cited by applicant]
AU 2022246430A1 · 2023 [cited by applicant]
WO WO2022079172A1 · 2022 [cited by applicant]
WO WO2022232783A1 · 2022 [cited by applicant]
WO WO2023288068A1 · 2023 [cited by examiner]
WO WO2023118551A1 · 2023 [cited by examiner]
Extended European Search Report and Written Opinion issued in European Patent Application No. 24184128.7 dated Dec. 9, 2024, in 08 pages. [cited by applicant]
Extended European Search Report and Written Opinion issued in European Patent Application No. 24184130.3 dated Nov. 11, 2024, in 07 pages. [cited by applicant]