Inferring moisture from color

Patent metadata

Field	Value
Publication number	US-12159457-B2
Application number	US-202318341434-A
Country	US
Kind code	B2
Filing date	Jun 26, 2023
Priority date	Jul 30, 2020
Publication date	Dec 3, 2024
Grant date	Dec 3, 2024

Abstract

Techniques are described herein for using artificial intelligence to predict crop yields based on observational crop data. A method includes: obtaining a first digital image of at least one plant; segmenting the first digital image of the at least one plant to identify at least one seedpod in the first digital image; for each of the at least one seedpod in the first digital image: determining a color of the seedpod; determining a number of seeds in the seedpod; inferring, using one or more machine learning models, a moisture content of the seedpod based on the color of the seedpod; and estimating, based on the moisture content of the seedpod and the number of seeds in the seedpod, a weight of the seedpod; and predicting a crop yield based on the moisture content and the weight of each of the at least one seedpod.

First claim

What is claimed is:

1. A method implemented by one or more processors, the method comprising:
receiving training data including a digital image of a live plant in a field, wherein the digital image is labeled based on a ground truth moisture content of a first seedpod of a first live plant;
generating preprocessed training data using the training data by:
segmenting the digital image to identify the first seedpod of the first live plant in the digital image;
determining a color of the first seedpod; and
determining a number of seeds in the first seedpod;

training one or more machine learning models to predict one or both of a moisture content of the first seedpod and a weight of the first seedpod based on the preprocessed training data and the ground truth moisture content; and
predicting a crop yield of a second seedpod of a second live plant based on a predicted moisture content of the first seedpod and a predicted weight of the first seedpod.

2. The method of claim 1 , wherein the generating the preprocessed training data further includes determining a size of each of the first seedpod in the digital image.

3. The method of claim 1 , wherein the weight of the first seedpod is a wet weight.

4. The method of claim 1 , wherein one or more of the machine learning models is a convolutional neural network model.

5. The method of claim 1 , wherein the digital image is obtained using a multi-camera array.

6. The method of claim 5 , wherein the digital image includes a plurality of digital images obtained at a plurality of positions along a length of a row of plants.

7. At least one non-transitory computer-readable storage medium comprising program instructions to cause at least one processor to:
receive training data including a digital image of a live plant, wherein the digital image is labeled based on a ground truth moisture content of a first seedpod of a first live plant;
generate preprocessed training data using the training data, wherein to generate the preprocessed training data includes to:
segment the digital image to identify the first seedpod of the first live plant in the digital image;
determine a color of the first seedpod; and
determine a number of seeds in the first seedpod;

train one or more machine learning models to predict one or both of a moisture content of the first seedpod and a weight of the first seedpod based on the preprocessed training data and the ground truth moisture content; and
predict a crop yield of a second seedpod of a second live plant based on a predicted moisture content of the first seedpod and a predicted weight of the first seedpod.

8. The at least one non-transitory computer-readable storage medium of claim 7 , wherein to generate the preprocessed training data further includes instructions to determine a size of each of the first seedpod in the digital image.

9. The at least one non-transitory computer-readable storage medium of claim 7 , wherein the weight of the first seedpod is a wet weight.

10. The at least one non-transitory computer-readable storage medium of claim 7 , wherein one or more of the machine learning models is a convolutional neural network model.

11. The at least one non-transitory computer-readable storage medium of claim 7 , wherein the digital image is obtained using a multi-camera array.

12. The at least one non-transitory computer-readable storage medium of claim 11 , wherein the digital image includes a plurality of digital images obtained at a plurality of positions along a length of a row of plants.

13. A system comprising:
a processor;
one or more computer-readable storage media; and
program instructions stored on the one or more computer-readable storage media, the program instructions to cause the processor to:
receive training data including a digital image of a live plant in a field, wherein the digital image is labeled based on a ground truth moisture content of a first seedpod of a first live plant;
generate preprocessed training data using the training data, wherein to generate the preprocessed training data using the training data includes to:
segment the digital image to identify the first seedpod of the first live plant in the digital image;
determine a color of the first seedpod; and
determine a number of seeds in the first seedpod;

train one or more machine learning models to predict one or both of a moisture content of the first seedpod and a weight of the first seedpod based on the preprocessed training data and the ground truth moisture content; and
predict a crop yield of a second seedpod of a second live plant based on a predicted moisture content of the first seedpod and a predicted weight of the first seedpod.

14. The system of claim 13 , wherein, to generate the preprocessed training data, the program instructions are to cause the processor to determine a size of each of the first seedpod in the digital image.

15. The system of claim 13 , wherein the weight of the first seedpod is a wet weight.

16. The system of claim 13 , wherein one or more of the machine learning models is a convolutional neural network model.

17. The system of claim 13 , wherein the digital image is obtained using a multi-camera array.

18. The system of claim 17 , wherein the digital image includes a plurality of digital images obtained at a plurality of positions along a length of a row of plants.

Assignees

• Deere & Co

Inventors

• Yuan Bodi
• Yuan Zhiqiang
• Zheng Ming

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06T7/90

  Determination of colour characteristics · CPC title

• A01G25/16

  Control of watering · CPC title