| CPC G06F 9/485 (2013.01) [G06F 9/542 (2013.01)] | 15 Claims |
|
1. A method for managing a computer vision environment, the method comprising:
identifying, by a computer vision (CV) node of a plurality of CV nodes, a CV alert, wherein the plurality of CV nodes are operably connected over a network, and at least one of the plurality of CV nodes comprises a physical camera;
in response to identifying the CV alert:
identifying candidate CV nodes of the plurality of CV nodes;
generating a dynamic CV zone using the candidate CV nodes;
making a determination that a CV node of the dynamic CV zone does not include a CV workload associated with the CV alert;
in response to the determination:
receiving a request, via computer-readable code, to instantiate the CV workload on the CV node of the dynamic CV zone;
in response to the request, instantiating the CV workload on the CV node of the dynamic CV zone;
initiating, in response to the instantiating, performance of the CV workload by the dynamic CV zone to generate CV data associated with the CV alert, wherein initiating the performance comprises:
requesting the CV node generate a first portion of the CV data using the CV workload,
making a determination that a second CV node of the dynamic CV zone does not include the CV workload, wherein the second CV node is one of the candidate CV nodes,
in response to the determination:
sending the CV workload to the second CV node, and
requesting the second CV node to generate a second portion of the CV data using the CV workload;
generating a CV alert case associated with the CV alert, wherein the CV alert case is a data structure;
obtaining the CV data from the dynamic CV zone;
updating the CV alert case using the CV data generated during the performance of the CV workload; and
providing the updated CV alert case to a video management system (VMS).
|