Edge-deployed automated target recognition and tracking system with distributed processing architecture

A computer system comprising:

• a hardware memory, wherein the computer system is configured to execute software instructions stored on non-transitory machine-readable storage media that:
• receive sensor data from an imaging system at an edge computing platform;
• process images through one or more machine learning models to detect objects within the sensor data;
• transform detected objects from image coordinates to geodetic coordinates using geometric transformations;
• maintain track states for the detected objects by associating detections across multiple time periods;
• monitor the track states to identify tracked objects approaching a coverage boundary of the imaging system;
• generate custody transfer messages for the identified objects, wherein the custody transfer messages include state information sufficient to enable continued tracking;
• transmit the custody transfer messages to one or more external tracking nodes before the identified objects exit coverage;
• receive incoming custody messages from the one or more external tracking nodes for objects entering coverage;
• and
• initialize tracking of objects identified in the incoming custody messages using the received state information.