無人飛行載具之應用-視覺追蹤

Visual Tracking for Unmanned Aerial Vehicle (UAV)

謝銘展(Ming-Jhan Shie)

指導教授 ： 鄭志宏

共同指導教授 ： 林義隆(Yih-Lon Lin)

義守大學/電機資訊學院/資訊工程學系/碩士(2017年)

摘要

近年來，無人飛行載具（Unmanned Aerial Vehicle，UAV）發展蓬勃，從早期的軍事應用，到現今已廣泛用於民生用途，將各式各樣的工作及任務交由無人飛行載具必為一股趨勢。無人飛行載具不須駕駛員，所以他能在環境較為險惡的地方執行飛行任務。在軍事上，無人飛行載具可用來偵查、監視、導航等用途。而在民間則能用來收集氣象資料、交通巡邏、噴灑農藥等任務。不過使用無人飛行載具執行任務時，還是需要駕駛員操控遙控器甚至還要有一名拍攝手來協助機上攝影機的控制及拍攝。
目前在飛機上的自動飛行主要都是依賴全球定位系統（Global Positioning System，GPS）不過這種感測器與影像資訊無關，必須事先知道任務目標的座標資訊才能進行設定，無法在飛行中自動辨認目標進行追蹤。所以本研究的目的則是將物件追蹤的技術應用在無人飛行載具上，由機上攝影機所拍攝的影像中自動搜尋一個已知顏色及形狀的特定目標。本研究使用HSV色彩空間描述目標之顏色特性，以顏色過濾符合條件的像素，之後再使用區塊偵測（Blob Detection）算出目標物之重心，再以遠端遙控或使用機載上樹莓派進行追蹤。其中研究範圍還有包含無人飛行載具與地面控制站以及影像傳輸之連結，使之能於地面控制站顯示資料及對無人飛行載具進行控制。

關鍵字

無人飛行載具 ； 多旋翼 ； 顏色擷取 ； 形態學 ； 區塊偵測 ； 物件追蹤

並列摘要

Unmanned Aerial Vehicle (UAV) has been developed for many years. Among the UAVs, the Quadcopters Take-Off and Landing develop rapidly. Quadcopters have developed into a steady flying unmanned vehicle. As robotics rapidly develops, many applications on robots to enhance its sensor and artificial intelligence can actually be applied to the quadcopter as well. The goal of this research is to enhance the video content applications, in which the videos are captured from the camera on the Quadcopter. The camera takes photographs and records video and in the meanwhile, we add the exceptional ability to track the objects by making use the color features.
In this study, we use Python to implement the color-tracking functionality. We introduce Ground Control Stations and the relative software of the UAV. The main autopilot is Ardupilot 3.3 from 3DR Robotics. We use the Open Source Computer Vision Library (OpenCV) to implement our color-tracking algorithm. Essentially, we compute the centroid of an object and program the UAV to track the centroid. In the image processing, we use binary segmentation on the component Hue in the HSV color space. Then we perform morphology open operation on the resulting binary image. At last, we use blob detection to compute the centroid of an object to track. In the UAV telemetry, we use Dronekit via MAVLink protocol to receive the status and send corresponding commands such as the altitude and attitudes of the UAV.

並列關鍵字

Unmanned Aerial Vehicle (UAV) ； Quadcopter ； morphology ； blob detection ； Object tracking ； Dronekit

參考文獻

[11]鄭光亨, “使用四旋翼直升機之顏色物件追蹤” 碩士論文, 義守大學, 2015

[1] W. Hu, T. Tan, L. Wang, and S. Maybank, “A survey on visual surveillance of object motion and behaviors,” IEEE Trans. Syst. Man Cybern. Part C Appl. Rev., vol. 34, no. 3, pp. 334–352, Aug. 2004.

[3] N. Paragios and R. Deriche, “Geodesic active contours and level sets for the detection and tracking of moving objects,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 22, no. 3, pp. 266–280, Mar. 2000.

[5] D. S. Jang and H. I. Choi, “Active models for tracking moving objects,” Pattern Recognit., vol. 33, no. 7, pp. 1135–1146, Jul. 2000.

[6] D. Comaniciu, V. Ramesh, and P. Meer, “Kernel-based object tracking,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, no. 5, pp. 564–577, May 2003.