I let AI deliver packages... it was a disaster

This is a simulated autonomous drone that locates a very small target, filters noisy detections, and drops a payload — all in real time.
In this video I walk through the ideas and code behind the system: a Unity drone that streams images to a server, real-time small-object detection using OpenVino as an optimizer, statistical noise cancellation using an Exponential Moving Average (EMA) and Evidence Accumulation, and a clean state-diagram decision flow that makes autonomous actions safe and predictable.

How EMA + evidence accumulation stabilizes noisy detections

System design: Unity client, server inference, and decision states

Short code highlights and architecture reasoning (not a line-by-line course)
====
Project Source Code: Github: https://github.com/zeyad-shaban

Music:
69 Bpm Bass rythm 3y9pi By Nate_the_late
Cat Walk By Sakura Girl

Видео I let AI deliver packages... it was a disaster канала ZeyadCode