在過去十年中，無人機在自主導航和航空攝影中的使用受到越來越多的關注。因為體積小、高解析度相機、智慧功能和移動性，無人機能夠輕易飛到人類難以到達的地方。在目前主流系統中，無人機玩家需要遠端控制無人機拍攝照片，這需要大量時間來反覆查看影像和控制調整無人機的位置和方向。因此，迫切需要開發一種自主飛行系統，以消除對無人機的手動控制，並在需要時主動有效地拍攝出符合偏好格式的自拍照。本研究的主要目標是開發一種基於視覺的自拍無人機系統，以使用手機應用程式以全自主的方式執行指定的自拍任務。
本研究報告了使用手機應用程式控制的基於視覺的全自主自拍系統的成功測試。任務從使用者在應用程式上選擇自拍範本開始，無人機將自動拍攝出與所選範本匹配的影像。該應用程式使用 DJI 軟體開發套件 (SDK) 的 android studio 軟體中設計的，稱為 MobileSDK。基於視覺的系統使用影像處理技術進行關鍵點提取，並部署了透視 n 點 (PnP) 演算法來估計校準相機的位置和方向。基於自拍範本的飛行控制用於引導無人機到目標位置以拍攝所需的影像。通過在室內和室外環境中進行的一系列實驗，展示了這種基於視覺的全自主自拍無人機系統的穩健性和效率。

The use of drones in autonomous navigation and aerial photography has received increased attention during the last decade. Due to small size, high-quality HD camera, intelligent features, and mobility, drones can easily reach places, which are difficult for humans to access. In prevailing systems, the user of the drone needs to remotely control the drone to a specific location to capture photographs of desired content formation. This requires a considerable amount of time on iteratively viewing images and tuning the position and orientation of the drone. Thus, there is a pressing demand to develop an autonomous flight system that will eliminate the use of manual controls of drones and capture selfie images meeting the preferred format both actively and efficiently when needed. The main objective of this study is to develop a vision-based selfie drone system to conduct the assigned selfie-mission in a fully autonomous manner using a cell phone application.
This research reports successful testing of a vision-based autonomous selfie system controlled using a cell phone application. The mission starts at the selection of the selfie template by the user on the app, and the drone will autonomously capture the image matching the selected template. The app has been designed in an Android studio software using DJI software development kit (SDK) called MobileSDK. The vision-based system uses image-processing techniques for keypoints extraction. The perspective-n-point (PnP) algorithm is deployed to estimate the position and orientation of a calibrated camera. Model-based flight control is used to guide the drone to the target position for capturing the required image. The robustness and efficiency of this vision-based autonomous selfie drone system are presented through a series of experiments conducted in indoor and outdoor environments.

[1] Cauchard, J. R., E, J. L., Zhai, K. Y., & Landay, J. A, “Drone & me: An Exploration Into Natural Human-Drone Interaction,” Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing - Ubicomp, 2015, doi:10.1145/2750858.2805823.
[2] G. Yee, "The 6 best foldable selfie drones to buy," drones fuel, 2017. [Online]. Available: https://dronesfuel.com/the-6-best-foldable-selfie-drones-to-buy-starting-from-50/. [Accessed July 2021].
[3] Craig J., “Best selfie drones to buy in 2021”, [Online]. Available: https://airdronecraze.com/selfie-drones/. [Accessed July 2021].
[4] Lukasbieri, “drone photos”, [Online]. Available: https://pixabay.com/nl/users/lukasbieri-4664461/. [Accessed July 2021].
[5] Yuncheng Lu, Zhucun Xue, Gui-Song Xia & Liangpei Zhang, “A survey on vision-based UAV navigation,” Geo-spatial Information Science, 21:1, 21-32, 2018, DOI: 10.1080/10095020.2017.1420509.
[6] S. Huang, Y. Chen, T. Yuan, S. Qi, Y. Zhu, and S.-C. Zhu, “Perspectivenet: 3d object detection from a single rgb image via perspective points,” in Advances in Neural Information Processing Systems, 2019, pp. 8905–8917.
[7] DJI, “Phantom 4,” [Online]. Available: https://www.dji.com/tw/. [Accessed July 2021].
[8] ZeroZero Robotics, “Your self-flying personal photographer”, [online]. Available: https://zerozero.tech/hover-camera-passport?d=pc. [Accessed July 2021].
[9] Zero-x, “ Evolved 4K”, [Online]. Available https://www.zero-x.com.au/product-evolved-4k. [Accessed July 2021].
[10] Coaguila, R.; Sukthankar, G.; and Sukthankar, R., “Selecting vantage points for an autonomous quadcopter videographer,” In Proceedings of Florida Artificial Intelligence Research Society, 386–391, 2016.
[11] S. Suzuki and Y. Mitsukura, "Mobile Robot Photographer," 2nd IAPR Asian Conference on Pattern Recognition, 2013, pp. 742-743, doi: 10.1109/ACPR.2013.192.
[12] M. Zabarauskas and S. Cameron, "Luke: An autonomous robot photographer," 2014 IEEE International Conference on Robotics and Automation (ICRA), 2014, pp. 1809-1815, doi: 10.1109/ICRA.2014.6907096.
[13] Morgan Quigley, Brian Gerkey, Ken Conley, Josh Faust, Tully Foote, Jeremy Leibs, Eric Berger, Rob wheeler, and Andrew Ng. ROS: an Open-Source Robot Operating System. In Proceedings of the 2009 IEEE International Conference on Robotics and Automation, 2009.
[14] G. J. Hsu and J. Huang, "A photographer robot with multiview face detector," 2016 IEEE International Conference on Industrial Technology (ICIT), 2016, pp. 2152-2156, doi: 10.1109/ICIT.2016.7475103.
[15] C. Lin and C. Liang, "Innovative Drone Selfie System and Implementation," 2017 International Symposium on Computer Science and Intelligent Controls (ISCSIC), 2017, pp. 135-140, doi: 10.1109/ISCSIC.2017.31.
[16] P.Corke, “Robotics, Vision and Control”, Springer Tracts in Advanced Robotics, 2011.
[17] Siddique, T.H., & Usman, M., “3D Object Localization Using 2D Estimates for Computer Vision Applications,”. ArXiv, abs/2009.11446. 2020.
[18] Roger Lin, “Camera frame model”, [Online]. Available: https://github.com/ptzu/abcd404 04.github.io/tree/master/2018/08/08/camera-frame-world-frame. [Accessed July 2021].
[19] Zhang, Zhengyou., “A Flexible New Technique for Camera Calibration. Pattern Analysis and Machine Intelligence,” IEEE Transactions on Pattern Analysis and Machine Intelligence. Doi:10.1109/34.888718, 2000.
[20] Aditi Majumder, “Camera Calibration”, [Online]. Available: https://www.ics.uci.edu/ ~majumder/VC/classes/cameracalib.pdf. [Accessed July 2021].
[21] Silverwind, “ Pinhole Camera: coordinate imaging principle”, [Online]. Available: https:// silverwind1982.pixnet.net/blog/post/134551091. [ Accessed: July 2021.
[22] Mathworks, “Camera models: camera calibration”, [Online]. Available: https:// www.mathworks.com/help/vision/ug/camera-calibration.html. [Accessed July 2021].
[23] Google, “Camera calibration”, [Online]. Available: https://docs.opencv.org/master/dc/ dbb/tutorial_py_calibration.html. [Accessed: July 2021].
[24] Leng, Dawei & Sun, WD, “Finding all the solutions of PnP problem,”. 2009 IEEE International Workshop on Imaging Systems and Techniques. doi: 10.1109/IST.2009.5071663. 2009
[25] A. Kaehler and G. Bradski, “Learning OpenCv3”, O’Reilly Media Inc. 2017.
[26] DJI Developer, “Flight Control”, [Online], Available: https://developer.dji.com/mobile-sdk/documentation/introduction/flightController_concepts.html. [Accessed: July 2021]
[27] Android Developer, “Android Studio User Guide”, [Online], Available: https://developer. android.com/studio/intro. [Accessed: July 2021].
[28] Mahesh Sharma, “Java Development Kit”, [Online], Available: https://medium.com/ maheshjtp/what-is-the-difference-between-jdk-jvm-jre-770f73c27c2f. [Accessed: July 2021].