Systems and methods for satellite image processing to estimate crop yield

Patent metadata

Field	Value
Publication number	US-9953241-B2
Application number	US-201514971610-A
Country	US
Kind code	B2
Filing date	Dec 16, 2015
Priority date	Dec 16, 2014
Publication date	Apr 24, 2018
Grant date	Apr 24, 2018

Abstract

Systems and methods for generating a crop yield estimate for an area as small as an individual field from images captured by a satellite are disclosed. The system generates simulations of crop yields in a region that includes the area by applying combinations of different parameters to a crop yield models. Observable quantities for simulated yields are determined from the simulations. The simulations and the observable properties are used to train a statistic model for the region that has two or more variables. Images captured by a satellite that include at least a portion of the area are obtained. Crop information is then determined from the images and weather information associated with the dates that the images where captured is obtained. The statistical model is then applied to the crop information and the weather information to determine a crop yield estimate.

First claim

What is claimed is:

1. A method for processing satellite images to determine a crop yield estimate for an area as small as an individual field in a region from images captured by a satellite comprising:
receiving a plurality of images captured by a satellite in a processing system, wherein each of the plurality of images includes pixels that correspond to an area and is captured on a particular date;
obtaining weather information for the area on the date that each of the plurality of images was captured using the processing system;
determining a plurality of observable quantities for the area from the pixels in the plurality of images that correspond to the area using the processing system;
determining a crop yield estimate for the area by applying a statistical model of the region including the area to the weather information obtained for the area on at least two different dates and the plurality of observable quantities for the area for the at least two different dates determined from the plurality of images using the processing system, wherein the statistical model has a plurality of variables including at least one variable from the plurality of observable quantities and at least one variable from the weather information and wherein the weather information for the area is of a same temporal scale as the weather information used to generate the statistical model; and
providing the crop yield estimate to an application for use in showing a predicted crop yield for the area.

2. The method of claim 1 wherein the statistical model is a multiple variable linear regression.

3. The method of claim 1 further comprising:
generating a plurality of crop yield simulations for the region using a processing system;
determining the plurality of observable quantities from each of the plurality of crop yield simulations using a processing system; and
training the statistical model for the region based upon the plurality of crop yield simulations and the plurality of observable quantities using the processing system.

4. The method of claim 3 wherein the plurality of crop yield simulations are generated using historical crop information for the region and historical environmental information for the region.

5. The method of claim 4 wherein the generation of the plurality of crop yield simulations comprises applying a plurality of different sets of parameters to a crop yield model using the processing system wherein the plurality of different sets of parameters represent different combinations of crop information and environmental information.

6. The method of claim 5 wherein the crop yield model is a model selected from the group consisting of: Agricultural Production Systems simulator (APSIM), Hybrid-Maize, and DSSAT.

7. The method of claim 5 wherein the plurality of different sets of parameters includes at least one parameter selected from the group consisting of: soil conditions, climate conditions, and management conditions for the region.

8. The method of claim 3 wherein the observable quantities from each of the plurality of crop yield simulations include at least one observable quantity selected from the group consisting of: optical-based Vegetation indexes (Vis) derived from Leaf Area Indexes (LAIs) and thermal measurements derived from water stress.

9. The method of claim 1 wherein the plurality of observable quantities are determined by applying image processing to the plurality of images.

10. An image processing system for determining a crop yield estimate for an area as small as an individual field in a region from images captured by a satellite comprising:
at least one processor;
memory accessible by each of the at least one processors; and
instructions stored in the memory that direct the at least one processors to:
receive a plurality of images captured of by a satellite wherein each of the plurality of images include pixels that correspond to the area and is captured on a particular date;
obtain weather information for the area on the date that each of the plurality of images was;
determine a plurality of observable quantities for the area from the pixels the plurality of images that correspond to the area;
determine a crop yield estimate for the area by applying a statistical model of the region including the area to the weather information and the plurality of observable quantities obtained for the area on at least two different dates wherein the statistical model has a plurality of variables including at least one variable from the plurality of observable quantities and at least one variable from the weather information and wherein the weather information for the area is of a same temporal scale as the weather information used to generate the statistical model; and
provide the crop yield estimate to an application for use in showing predicted crop yield for the area.

11. The system of claim 10 wherein the statistical model is a multiple variable linear regression.

12. The system of claim 10 wherein the instructions further direct the at least one processors to:
generate a plurality of crop yield simulations for the region;
determine the plurality of observable quantities from each of the plurality of crop yield simulations; and
train the statistical model for the region based upon the plurality of crop yield simulations and the plurality of observable quantities.

13. The system of claim 12 wherein the plurality of crop yield simulations are generated using historical crop information for the region and historical environmental information for the region.

14. The system of claim 13 wherein the generation of the plurality of crop yield simulations is performed by applying a plurality of different sets of parameters to a crop yield model wherein the plurality of different sets of parameters represent different combinations of crop information and environmental information.

15. The system of claim 14 wherein the crop yield model is a model selected from the group consisting of: Agricultural Production Systems simulator (APSIM), Hybrid-Maize, and DSSAT.

16. The system of claim 14 wherein the plurality of different sets of parameters include at least one parameter selected from the group consisting of: soil conditions, climate conditions, and management conditions for the region.

17. The system of claim 12 wherein the observable quantities from each of the plurality of crop yield simulations include at least one observable quantity selected from the group consisting of: optical-based Vegetation indexes (Vis) derived from Leaf Area Indexes (LAIs) and thermal measurements derived from water stress.

18. A non-transitory machine readable medium accessible by one or more processor that store instructions that direct the one or more processors to perform the method comprising:
receiving a plurality of images captured of by a satellite in wherein each of the plurality of images includes pixels that correspond to a portion of an area in a region and is captured on a particular date;
obtaining weather information for the area on the date each image was captured;
determining a plurality of observable quantities for the area from the plurality of images;
determining a crop yield for the area by applying the statistical model of the region including the area to the weather information on at least two different data and the plurality of observable quantities on the two different dates determined from the plurality of images wherein the statistical model has a plurality of variables including at least one variable from the plurality of obs

Assignees

• Univ Leland Stanford Junior

Inventors

• Lobell David

Classifications

• G06T7/00 Primary

  Image analysis · CPC title

• G06V20/13

  Satellite images · CPC title

• G06T2207/20076

  Probabilistic image processing · CPC title

• G06K2009/00644

  Physics · mapped topic

• G06T2207/10036

  Multispectral image; Hyperspectral image · CPC title