嵌入式影像辨識系統應用於無人直昇機自動降落系統之開發

An Embedded Computer Vision System Application in the Automatic Landing System for Unmanned Aerial Helicopter

陳盈佐(Ying-Zuo Chen)

指導教授 ： 曾百由

國立臺北科技大學/機電學院/機電整合研究所/碩士(2012年)

摘要

為達到無人直昇機的自動降落，前人使用PC based平台，利用影像處理方法來辨識降落圖示；但隨著無人載具應用漸趨廣泛，對於無人載具的輕量化、系統穩定性以及電池續航力有了更高的要求。而嵌入式系統因體積小、處理速度佳、耗電量低以及具有穩定處理特定任務等特性，故本研究採用嵌入式系統為架構，並以影像識別為實現手段開發演算法，來達到辨識降落圖示之目的。
本研究系統架構以CCD攝影機為影像輸入裝置，嵌入式控制器「ADSP-BF561」為系統核心。藉由自行設計的降落圖示，以其顏色與幾何形狀做為識別特徵，完成影像辨識。
實驗結果顯示本嵌入式影像識別系統能成功辨識降落圖示，並能即時處理每秒30張的影像資料(33毫秒/張)。證明利用嵌入式影像辨識系統可以達成無人載具即時影像輔助降落的可行性。

關鍵字

無人載具 ； 自動降落 ； 嵌入式系統 ； 機器視覺

並列摘要

To complete auto landing for Vertical Take-Off and Landing(VTOL) unmanned aerial vehicle (UAV), earlier studies used PC based platform with image processing algorithms to identify and locate the landing pad; but as the VTOL UAV applications grow, specifications for its lightweight, system reliability and power consumption are more strict. And because the advantages of embedded system has small, good computing speed, less power consumption and reliable for specific processing etc, embedded system with image processing algorithms has been chosen to develop our landing system. This system has a wireless CCD camera as image input source, with ADSP-BF561 as the core of embedded system. Especially, the landing pad is defined with its unique color and shape arrangement to achieve image recognition.
Experiment results show that the embedded computer vision system can identify the defined landing pad, with the ability of processing image data flow at the rate of 30Hz speed(33 milli-seconds per frame). These results prove the feasibility of applying the designed embedded image processing system.

並列關鍵字

unmanned aerial vehicle ； automatic landing ； embedded system ； computer vision

參考文獻

[1]M. Bosse, “A vision-augmented navigation system for an autonomous helicopter,” Master’s thesis, Boston Univ., Boston, MA, 1997.

[2]O. Amidi, T. Kanade and K. Fujita, “A visual odometer for autonomous helicopter flight,” Robotics and Autonomous Systems, vol. 28, no. 2, pp.185-193, 1999.

[3]O. Shakernia, Y. Ma, T. J. Koo, and S. S. Sastry, “Landing an unmanned air vehicle: Vision-based motion estimation and nonlinear control,” Asian Journal of Control, vol. 1, pp.128-145, 1999.

[4]U.S. Department of Transportation Federal Aviation Administration, Advisory Circular-HELIPORT DESIGN, 2004.

[5]Wen-Hsiang Tsai and Zhi-Fang Yang, “Using parallel line information for vision-based landmark location estimation and an application to automatic helicopter landing,” Robotics and Computer Integrated Manufacturing, vol. 14,