競賽 · 邊緣運算 · 2023Competition · Edge Computing · 2023
邊緣運算無人機Edge-Computing UAV
自組四旋翼無人機搭載 Raspberry Pi 與相機，於機上進行 YOLO 影像辨識並即時回傳雲端。A self-built quadrotor with Raspberry Pi + camera running on-board YOLO inference and streaming to the cloud.
概述Overview
以實驗室自組四旋翼無人機為主體，搭載 Raspberry Pi 與相機模組，透過邊緣運算在機上以 YOLO 進行影像辨識（搜尋指定目標），並將空中視角即時傳輸至雲端資料庫供監控端取用。以 Mission Planner（地面站軟體）進行航點設置與飛行參數調整，並以 3D 列印（PLA／PET）製作最符合需求的客製化機架零件。本作品為「2023 大專校院機電智慧暨電動載具創意實作競賽」參賽作品。 Built on a lab-assembled quadrotor carrying a Raspberry Pi and camera, it runs edge inference with YOLO on-board (searching for a designated target) and streams the aerial view to a cloud database for a monitoring station. Waypoints and flight parameters are set in Mission Planner, and custom airframe parts are made by 3D printing (PLA/PET). It was an entry in the 2023 National EV & Mechatronics Innovation Contest.
重點Highlights
- 機上邊緣運算：Raspberry Pi 4 B + Camera Module 3，於空中即時推論。On-board edge computing: Raspberry Pi 4 B + Camera Module 3 for in-flight inference.
- YOLO 目標偵測（搜尋指定目標）。YOLO target detection (searching for a designated object).
- 空中影像即時回傳雲端，供監控端取用。Aerial imagery streamed to the cloud in real time for monitoring.
- PIXHAWK 2 飛控 + Mission Planner 航點與飛行參數設定。PIXHAWK 2 flight controller + Mission Planner for waypoints and flight parameters.
- 3D 列印（PLA／PET）客製化機架零件。3D-printed (PLA/PET) custom airframe parts.
競賽資訊Competition
- 競賽全名Event
- 2023 大專校院機電智慧暨電動載具創意實作競賽2023 National Collegiate Mechatronics & EV Innovation Contest
- 作品名稱Title
- 結合邊緣運算之無人機開發與應用UAV development & applications with edge computing
- 隊伍Team
- 高雄肝鐵人（潘彥良、吳維文、張勝冠、李主恩、莊鈞翔）"Kaohsiung Iron Team" (5 members)
- 指導老師Advisors
- 柯春旭、李青旻Chun-Hsu Ko, Ching-Min Lee
- 主辦Organizer
- 澄德科技教育基金會Chengde Technology Education Foundation
- 日期Date
- 2023/07/25