CPC G06T 7/80 (2017.01) [G06T 7/70 (2017.01); G06V 20/52 (2022.01); H04N 23/64 (2023.01); H04N 23/661 (2023.01); H04N 23/80 (2023.01); G06T 2207/30264 (2013.01)] | 14 Claims |
1. A method of operating a camera-based parking management system, comprising:
activating an AI-enabled camera covering a staging area responsive to detecting a vehicle;
capturing a first snapshot of the staging area using the AI-enabled camera;
transmitting the first snapshot to an edge processor coupled to the AI-enabled camera to identify one or more objects shown in the first snapshot;
determining that first object observed in the first snapshot is not capable of being identified with a predefined degree of certainty based on a machine learning model;
responsive to the determining:
identifying first one or more characteristics about the environmental conditions surrounding the first object that potentially reduce an identification accuracy of one or more vehicle-specific parameters; and
identifying second one or more characteristics about the first snapshot that potentially reduce the identification accuracy of one or more vehicle-specific parameters; and
automatically adjusting one or more settings of the AI-enabled camera to capture a second snapshot based on identifying the first one or more characteristics about the environmental conditions surrounding the first object;
automatically adjusting one or more settings of the processing of images captured by the AI-enabled camera based on identifying the second one or more characteristics about the first snapshot;
wherein the second one or more characteristics about the first snapshot include a resolution of the snapshot, network availability based on the time of day, and a network speed of a network connection between the AI-enabled camera and the edge processor;
capturing the second snapshot of the staging area using the AI-enabled camera;
determining the first object observed in the first snapshot is capable of being identified with the predefined degree of certainty based on the machine learning model; and
adjusting the machine learning model based on the adjustment of the settings of the AI-enabled camera and settings of the processing of images captured by the AI-enabled camera.
|