Back to results
US 11,237,299 B2Granted
Self-learning nowcast system for modeling, recording, and predicting convective weather
AssigneeI.M. Systems Group, Inc.
InventorRuiyue Chen
Life of this patent
Estimating the term…
Abstract
The systems, methods, and apparatuses described herein provide integrated weather forecast products designed to assist operations managers with operational decision-making related to a designated event or set of events. The present disclosure provides a way to process weather data from various sources and in diverse data formats containing varying spatial resolutions and temporal resolutions, in order to generate an integrated and cohesive weather projection product such that the weather projection product is continuous in both spatial and temporal domains relative to a designated event or set of events.
Claims38
A computer-implemented method for automated modeling, recording, and predicting convective weather, the method comprising:
- receiving, by a computer from one or more image data sources, one or more weather data digital pixel images for a set of time step intervals, the one or more weather data digital pixel images including at least a first weather data digital pixel image for a first time step interval;
- identifying, by the computer, an ongoing convective object in the first weather data digital pixel image based on the presence of one or more predetermined parameters that indicate the presence of a convective object;
- generating, by the computer, a convective object data record for the ongoing convective object, the convective object data record including one or more parameters associated with the ongoing convective object at the first time step interval and at one or more time step intervals after the first time step interval;
- storing, by the computer, the convective object data record in a convective evolution database;
- generating, by the computer, a nowcast output for the ongoing convective object based on (i) the convective object data record for the ongoing convective object and (ii) one or more convective object data records for one or more past convective objects, wherein the nowcast output is generated at least in part using a machine-learning nowcasting algorithm configured to generate a predictive nowcast of an ongoing convective object based on convective object data records for the ongoing convective object and one or more past convective objects;
- and
- transmitting, by the computer, the nowcast output to one or more user devices configured to display the nowcast output, wherein when displayed on the one or more user devices via a graphical user interface, the nowcast output includes a geographic map indicating a nowcast weather image of the ongoing convective object at a set of future time step intervals.
4 more elements
A weather nowcasting system for automated modeling, recording, and predicting convective weather, the system comprising:
- a convective evolution database comprising non-transitory machine-readable medium configured to store one or more convective object data records;
- and
- a processor configured to:
- receive, from one or more image data sources, one or more weather data digital pixel images for a set of time step intervals, the one or more weather data digital pixel images including at least a first weather data digital pixel image for a first time step interval;
- identify an ongoing convective object in the first weather data digital pixel image based on the presence of one or more predetermined parameters that indicate the presence of a convective object;
- generate a convective object data record for the ongoing convective object, the convective object data record including one or more parameters associated with the ongoing convective object at the first time step interval and at one or more time step intervals after the first time step interval;
- store the convective object data record in the convective evolution database;
- generate a nowcast output for the ongoing convective object based on (i) the convective object data record for the ongoing convective object and (ii) one or more convective object data records for one or more past convective objects, wherein the nowcast output is generated at least in part using a machine-learning nowcasting algorithm configured to generate a predictive nowcast of an ongoing convective object based on convective object data records for the ongoing convective object and one or more past convective objects;
- and
- transmit the nowcast output to one or more user devices configured to display the nowcast output, wherein when displayed on the one or more user devices via a graphical user interface, the nowcast output includes a geographic map indicating a nowcast weather image of the ongoing convective object at a set of future time step intervals.
7 more elements
Dependencies are inferred from the claim text.