Estimating soil properties within a field using hyperspectral remote sensing
A method is provided for determining soil properties for an area of land, from soil spectrum data. In an embodiment, the method includes receiving soil spectrum data records from hyperspectral sensors that represent a mean soil spectrum of a specific geo-location of the area of land and removing interference signals from the spectrum data records to create soil spectral bands. The method also includes predicting a plurality of soil property datasets based on a partial least-square regression and the soil spectral bands and selecting specific soil property datasets from the plurality of soil property datasets to represent soil properties of the specific geo-location, wherein the specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data. The specific soil property datasets may then be used to generate a crop prescription of recommended hybrid seeds or population densities for the specific geo-location.
1. A computer-implemented method comprising:
receiving, by a server computer system, 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;
removing, by the server computer system, interference signals from the one or more soil spectrum data records to create one or more soil spectral bands, wherein the removing includes one or more of data smoothing, application of spectral derivatives, standard normal variate preprocessing, and converting spectral reflectance to absorbance;
selecting, by the server computer system, a subset of the one or more soil spectral bands;
calculating, by the server computer system, a partial least-square regression between a first matrix and a second matrix, the first matrix including an eigen-decomposition of the subset of the one or more soil spectral bands of the one or more soil spectrum data records, the second matrix including soil property values;
predicting, by the server computer system, a plurality of soil property datasets based on the partial least-square regression;
selecting, by the server computer system, 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 one or more specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data; and
transmitting, by the server computer system, 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 for the specific geo-location of the area of land.
2. The computer-implemented method of claim 1 , wherein predicting the plurality of soil property datasets includes:
determining, by the server computer system, one or more latent variables from the partial least-square regression; and
predicting values of the plurality of soil property datasets further based on the one or more latent variables.
3. The computer-implemented method of claim 1 , wherein the partial least-square regression includes:
X=TP T +E
Y=UQ T +F
U=TB,
wherein X is the first matrix; wherein Y is the second matrix; wherein P and Q are loading matrices; wherein B is a regression coefficient; wherein E and F are residual matrices; and wherein T and U are matrices including datasets of potential latent variables.
4. The computer-implemented method of claim 1 , wherein the hyperspectral sensors are affixed to aerial equipment, and wherein receiving the one or more soil spectrum data records includes receiving the one or more soil spectrum data records from the hyperspectral sensors affixed to the aerial equipment.
5. The computer-implemented method of claim 1 , wherein the hyperspectral sensors are affixed to movable or stationary land equipment, and wherein receiving the one or more soil spectrum data records includes receiving the one or more soil spectrum data records from the hyperspectral sensors that are affixed to the movable or stationary land equipment.
6. The computer-implemented method of claim 1 , wherein removing the interference signals further 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.
7. The computer-implemented method of claim 6 , wherein removing the interference signals further comprises calculating a derivative of each moving average over a specified band distance.
8. The computer-implemented method of claim 1 , wherein predicting the plurality of 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; and
wherein the one or more locations are determined using spatial sampling of the one or more soil spectral bands.
9. The computer-implemented method of claim 1 , wherein the interference signals include at least one of a baseline drift effect, particle deviation, and/or surface heterogeneity.
10. One or more non-transitory storage media comprising executable instructions which, when executed by one or more computing devices, cause the one or more computing devices to perform steps of:
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;
removing interference signals from the one or more soil spectrum data records to create one or more soil spectral bands, wherein the removing includes one or more of data smoothing, application of spectral derivatives, standard normal variate preprocessing, and converting spectral reflectance to absorbance;
selecting a subset of the one or more soil spectral bands;
calculating a partial least-square regression between a first matrix and a second matrix, the first matrix including an eigen-decomposition of the subset of the one or more soil spectral bands of the one or more soil spectrum data records, the second matrix including soil property values;
predicting a plurality of soil property datasets based on the partial least-square regression;
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 one or more specific soil property datasets include property data and spectral band data for spectral bands used to determine the property data; and
transmitting 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.
11. The one or more non-transitory storage media of claim 10 , wherein predicting the plurality of soil property datasets includes:
determining one or more latent variables from the partial least-square regression; and
predicting values of the plurality of soil property datasets further based on the one or more latent variables.
12. The one or more non-transitory storage media of claim 10 , wherein the partial least-square regression includes:
X=TP T +E
Y=UQ T +F
U=TB,
wherein X is the first matrix; wherein Y is the second matrix; wherein P and Q are loading matrices; wherein B is a regression coefficient; wherein E and F are residual matrices; and wherein T and U are matrices including datasets of potential latent variables.
13. The one or more non-transitory storage media of claim 10 , wherein the hyperspectral sensors are affixed to aerial equipment, and wherein receiving the one or more soil spectrum data records includes receiving the one or more soil spectrum data records from the hyperspectral sensors that are affixed to the aerial equipment.
14. The one or more non-transitory storage media of claim 10 , wherein the hyperspectral sensors are affixed to movable or stationary land equipment, and wherein receiving the one or more soil spectrum data records includes receiving the one or more soil spectrum data records from the hyperspectral sensors that are affixed to the movable or stationary land equipment.
15. The one or more non-transitory storage media of claim 10 , wherein removing the 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.
16. The one or more non-transitory storage media of claim 15 , wherein removing the interference signals further comprises calculating a derivative of each moving average over a specified band distance.
17. The one or more non-transitory storage media of claim 10 , wherein predicting the plurality of 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; and
wherein the one or more locations are determined using spatial sampling of the one or more soil spectral bands.
18. The one or more non-transitory storage media of claim 10 , wherein the interference signals include at least one of a baseline drift effect, particle deviation, and/or surface heterogeneity.