Estimating soil properties within a field using hyperspectral remote sensing

Patent metadata

Field	Value
Publication number	US-10338272-B2
Application number	US-201514866160-A
Country	US
Kind code	B2
Filing date	Sep 25, 2015
Priority date	Sep 12, 2014
Publication date	Jul 2, 2019
Grant date	Jul 2, 2019

Abstract

A method for estimating soil properties within a field using hyperspectral remotely sensed data is provided. In an embodiment, estimating soil properties may be accomplished using a server computer system that receives, via a network, soil spectrum data records that are used to predict soil properties for a specific geo-location. Within the server computer system a soil preprocessing module receives one or more soil spectrum data records that represent a mean soil spectrum of a specific geo-location of a specified area of land. The soil preprocessing module then removes interference signals from the soil spectrum data, creating a set of one or more spectral bands. By removing interference signals, the spectral bands are not erroneously skewed from effects such as baseline drift, particle deviation, and surface heterogeneity. A soil regression module inputs the one or more soil spectral bands and predicts soil property datasets. The soil property datasets include specific soil properties relevant to determining fertility of the soil or soil property levels that may influence soil management at a specific geo-location. The soil regression module then takes the multiple soil property datasets and selects multiple specific soil property datasets that best represent the existing soil properties. Included in the soil property datasets are the multiple soil properties predicted and the spectral band data used to determine the specific soil properties. The soil regression module sends this predicted data to a soil model database.

First claim

What is claimed is:

1. A computer-implemented method comprising:
using a soil preprocessing module in a server computer system, receiving one or more soil spectrum data records from hyperspectral sensors that represent a mean soil spectrum of a specific geo-location of a specified area of land;
using the soil preprocessing module, removing interference signals from the one or more soil spectrum data records to create one or more soil spectral bands, wherein the interference signals include at least one of a baseline drift effect, particle deviation, and surface heterogeneity;
using a soil regression module, predicting a plurality of soil property datasets based upon the one or more soil spectral bands;
using the soil regression module, selecting one or more specific soil property datasets from the plurality of soil property datasets to represent soil properties of the specific geo-location, based on a quality score, wherein the specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data
using the soil regression module, sending the one or more specific soil property datasets to a soil database repository for generating a crop prescription that includes a recommended hybrid seed line or population density.

2. The method of claim 1 , comprising receiving the one or more soil spectrum data records from airborne hyperspectral sensors that are affixed to aerial equipment.

3. The method of claim 1 , comprising receiving the one or more soil spectrum data records from hyperspectral sensors that are affixed to movable land equipment.

4. The method of claim 1 , comprising receiving the one or more soil spectrum data records from hyperspectral sensors that are affixed to stationary equipment.

5. The method of claim 1 , wherein removing interference signals comprises calculating a set of moving averages from one or more subsets of the one or more soil spectrum data records, wherein each moving average is a sum of a subset of adjacent soil spectrum records multiplied by a calculated convolution coefficient.

6. The method of claim 5 , wherein removing interference signals further comprises calculating a derivative of each moving average over a specified band distance.

7. The method of claim 5 , wherein removing interference signals further comprises calculating a second derivative of each moving average over a specified band distance.

8. The method of claim 1 , wherein removing interference signals further comprises calculating a standard normal variate at a specified spectral band for the one or more soil spectrum data records, wherein the standard normal variate at a specified spectral band is a difference of a raw spectral value and an averaged spectral value over a set sample spectrum divided by a standard deviation of the set sample spectrum.

9. The method of claim 1 , wherein removing interference signals further comprises calculating absorbance values of the one or more soil spectrum data records, wherein the absorbance values equal a logarithmic function of an inverse of the one or more soil spectrum data records.

10. The method of claim 1 , wherein predicting soil property datasets based upon one or more soil spectral bands further comprises:
using a spectral configuration module, configuring a band selection module to use a specified set of soil spectral bands for selecting a subset of soil spectral bands;
using the band selection module, selecting the subset of soil spectral bands for soil property evaluation; and
using the soil regression module, predicting soil property datasets based upon the subset of soil spectral bands.

11. The method of claim 10 , wherein selecting a subset of soil spectral bands comprises:
selecting a population set of randomly generated spectral band combinations;
generating an offspring set by exchanging properties from the population set of randomly generated spectral band combinations;
generating a mutation set by altering properties of each of the offspring set in order to simulate random disturbance;
selecting a subset of soil spectral bands from the mutation set, wherein the selection is based upon a preconfigured set of soil spectral bands.

12. The method of claim 1 , wherein predicting soil property datasets comprises:
receiving soil property data based upon one or more ground soil samples from one or more locations identified within the specific geo-location, wherein the one or more locations are determined using spatial sampling of the one or more soil spectral bands.

13. The method of claim 1 , wherein predicting soil property datasets comprises:
calculating a partial least square regression set between a first matrix and a second matrix;
wherein the first matrix comprises eigen-decompositions of spectral values of the one or more soil spectrum data records and the second matrix comprises soil properties;
determining one or more latent variables from the partial least square regression set;
creating soil property datasets using the one or more latent variables to predict the values in the soil property datasets.

14. The method of claim 1 , wherein selecting the one or more soil property datasets comprises:
calculating the quality score for the soil property datasets;
wherein the quality score is a root mean squared error;
wherein the squared error is a sum difference between predicted values of the soil property datasets and observed values of the soil property datasets.

15. The method of claim 1 , further comprising:
using the soil regression module, sending the one or more specific soil property datasets to a spectral configuration module and configuring a band selection module using the specific soil property dataset.

16. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of a computer-implemented method comprising the steps of:
using a soil preprocessing module in a server computer system, receiving one or more soil spectrum data records from hyperspectral sensors that represent a mean soil spectrum of a specific geo-location of a specified area of land;
using the soil preprocessing module, removing interference signals from the one or more soil spectrum data records to create one or more soil spectral bands, wherein the interference signals include at least one of a baseline drift effect, particle deviation, and surface heterogeneity;
using a soil regression module, predicting a plurality of soil property datasets based upon the one or more soil spectral bands;
using the soil regression module, selecting one or more specific soil property datasets from the plurality of soil property datasets to represent soil properties of the specific geo-location, based on a quality score, wherein the specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data;
using the soil regression module, sending the one or more specific soil property datasets to a soil database repository for generating a crop prescription that includes a recommended hybrid seed line or population density.

17. The one or more non-transitory storage media of claim 16 , comprising receiving the one or more soil spectrum data records from airborne hyperspectral sensors that are affixed to aerial equipment.

18. The one or more non-transitory storage media of claim 16 , comprising receiving the one or more soil spectrum data records from hyperspectral sensors that are affixed to movable land equipment.

Assignees

• Climate Corp

Inventors

• Xiang Haitao
• Yang Xianyuan
• Koshnick Nick
• Cisek Nick

Classifications

• E02D1/04 Primary

  Sampling of soil {(E02D1/025 takes precedence)} · CPC title

• G16C20/30

  Prediction of properties of chemical compounds, compositions or mixtures · CPC title

• G16C20/20

  Identification of molecular entities, parts thereof or of chemical compositions · CPC title

• G01N21/17

  Systems in which incident light is modified in accordance with the properties of the material investigated (where the material investigated is optically excited causing a change in wavelength of the incident light G01N21/63) · CPC title

• G16C99/00

  Subject matter not provided for in other groups of this subclass · CPC title