US 12,461,273 B2
Vehicle-based anomaly detection using artificial intelligence and combined environmental and geophysical sensor data
Daniel P. Lathrop, University Park, MD (US); Heidi Myers, Greenbelt, MD (US); and Vedran Lekic, Washington, DC (US)
Assigned to University of Maryland, College Park, College Park, MD (US)
Filed by University of Maryland, College Park, College Park, MD (US)
Filed on Dec. 15, 2021, as Appl. No. 17/552,175.
Claims priority of provisional application 63/126,485, filed on Dec. 16, 2020.
Prior Publication US 2024/0134085 A1, Apr. 25, 2024
Int. Cl. G01V 11/00 (2006.01); G06N 3/04 (2023.01)
CPC G01V 11/00 (2013.01) [G06N 3/04 (2013.01)] 43 Claims
OG exemplary drawing
 
1. A method for producing a map depicting geophysical anomalies in an area based on geophysical sensor data, the method comprising:
(a) accessing environmental sensor data with a computer system, the environmental sensor data being indicative of environmental conditions in an area;
(b) accessing a first machine learning algorithm with the computer system, wherein the first machine learning algorithm has been trained on first training data to determine confidence weights from environmental conditions;
(c) generating confidence weight values by applying the environmental sensor data to the first machine learning algorithm, generating a first output as the confidence weight values;
(d) accessing geophysical sensor data with a computer system, the geophysical sensor data being indicative of data acquired from the area using a plurality of different geophysical sensors;
(e) accessing a second machine learning algorithm with the computer system, wherein the second machine learning algorithm has been trained on second training data to identify anomalies in the area from the geophysical sensor data acquired from the area;
(f) generating anomaly map data by applying the geophysical sensor data to the second machine learning algorithm and using the confidence weight values as an input to the second machine learning algorithm to control contributions of different ones of the geophysical sensors when generating the anomaly map data, generating a second output as the anomaly map data; and
(g) providing the anomaly map data to a user, wherein the anomaly map data comprise an anomaly map depicting at least one anomaly in the area.