US 12,245,536 B2
Systems, methods, and apparatuses for implementing automated data modeling and scaling of a soil health data fabric
Todd Taylor, Mesa, AZ (US); Matthew Ikle, Alamosa, CO (US); and William Brandt, Phoenix, AZ (US)
Assigned to Arizona Board of Regents on Behalf of Arizona State University, Scottsdale, AZ (US)
Filed by Arizona Board of Regents on behalf of Arizona State University, Scottsdale, AZ (US)
Filed on Oct. 26, 2021, as Appl. No. 17/511,288.
Claims priority of provisional application 63/105,740, filed on Oct. 26, 2020.
Prior Publication US 2022/0124963 A1, Apr. 28, 2022
Int. Cl. A01B 79/00 (2006.01); G06F 18/214 (2023.01); G06V 20/10 (2022.01); H04L 9/32 (2006.01); H04L 9/00 (2022.01)
CPC A01B 79/005 (2013.01) [G06F 18/2148 (2023.01); G06V 20/188 (2022.01); H04L 9/3236 (2013.01); H04L 9/50 (2022.05)] 20 Claims
OG exemplary drawing
 
1. A system including a convolutional neural network to predict soil quality estimates for a target region using data measurements for a soil health data fabric, the system comprising:
a memory configured to store the data measurements for the soil health data fabric; and
processing circuitry in communication with the memory, wherein the processing circuitry is configured to:
compile the data measurements from disparate soil health data sources at a plurality of training data sites into a blockchain repository;
combine, using a track and trace function of the blockchain repository, the data measurements with inputs including at least a plurality of baseline soil samples, and satellite imagery related to the plurality of training data sites to correlate soil health;
train, by the system, the convolutional neural network, based on the data measurements combined with the inputs using a selected machine learning algorithm to generate a trained artificial intelligence model, wherein the selected machine learning algorithm includes transfer learning and feedback processing;
execute the selected machine learning algorithm to update the trained artificial intelligence model using the transfer learning to generate a plurality of local soil quality estimates based on determined local soil quality estimates transferred from the data measurements combined with the inputs, wherein the plurality of local soil quality estimates are based on pre-determined soil quality indicators;
update the local soil quality estimates of the trained artificial intelligence model using the feedback processing to generate a regional soil quality estimate, wherein to update the local soil quality estimates of the trained artificial intelligence model includes application of the transfer learning and the feedback processing to update the trained artificial intelligence model using additional data in comparison to the plurality of baseline soil samples;
scale, by the trained artificial intelligence model, the local soil quality estimates and the regional soil quality estimate; and
output, by the trained artificial intelligence model, the soil quality estimates for the target region based on extended satellite imagery.