SecureDrop Navigator: Image based Delivery Classifier

In the realm of autonomous drone package delivery, evaluating the safety of designated drop zones has been a relatively understudied aspect. Ensuring the secure nature of these zones before and during the delivery process is crucial for overall safety. This paper introduces a novel and efficient classifier designed to assess the safety of a specified drop zone utilizing a single onboard camera. This classifier is the first of its kind to address safety evaluation precisely at the point of drone delivery. Through experimentation on recorded drone videos, the proposed classifier demonstrates an impressive average precision and average recall of across various test scenarios.