US 11,704,576 B1
Identifying ground types from interpolated covariates
John A. McEntire, Park City, UT (US); and Thomas A. Dye, Austin, TX (US)
Assigned to ARVA INTELLIGENCE CORP., Salt Lake City, UT (US)
Filed by ARVA INTELLIGENCE CORP., Park City, UT (US)
Filed on Feb. 19, 2021, as Appl. No. 17/180,695.
Application 17/180,695 is a continuation in part of application No. 17/171,887, filed on Feb. 9, 2021.
Application 17/171,887 is a continuation in part of application No. 17/160,286, filed on Jan. 27, 2021.
Claims priority of provisional application 63/100,545, filed on Mar. 17, 2020.
Claims priority of provisional application 62/995,948, filed on Feb. 20, 2020.
Claims priority of provisional application 62/995,674, filed on Feb. 7, 2020.
Claims priority of provisional application 62/995,484, filed on Jan. 29, 2020.
Int. Cl. G06F 11/30 (2006.01); G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06Q 50/02 (2012.01); G06F 16/29 (2019.01); G06F 18/214 (2023.01); G06F 18/21 (2023.01); G06F 18/2413 (2023.01); G06F 18/243 (2023.01)
CPC G06N 5/01 (2023.01) [G06F 16/29 (2019.01); G06F 18/214 (2023.01); G06F 18/217 (2023.01); G06F 18/2414 (2023.01); G06F 18/24323 (2023.01); G06N 20/00 (2019.01); G06Q 50/02 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for identifying ground types from one or more interpolated covariates, the method comprising:
providing a geospatial database associated with a crop prediction engine, wherein the geospatial database receives a plurality of soil composition information for each of a plurality of plots of land;
accessing the plurality of soil composition information for each of the plurality of plots of land, in which the soil composition information includes at least one of a plurality of measured soil sample results, a plurality of environmental results, and a plurality of soil conductivity results;
identifying a plurality of covariates from the plurality of soil composition information;
interpolating, at the crop prediction engine, covariates associated with a plurality of different locations with an interpolation training model;
generating a plurality of voxels, wherein each voxel associates the one or more interpolated covariates with a corresponding geographical location within at least one of the plurality of plots of land;
training a random forest training model with the interpolated covariates;
traversing the voxels through the trained random forest model to identify one or more clusters of voxels that are co-associated; and
identifying a ground type by combining the one or more co-associated clusters, wherein each ground type is associated with at least one of a crop zone, a soil fertility, and a farm management recommendation.