研究生: |
林郁程 Yu-Cheng Lin |
---|---|
論文名稱: |
撞球辨識深度學習與撞球策略分析 Deep learning in billiard balls detection and analysis of billiard strategy |
指導教授: |
施慶隆
Ching-Long Shih |
口試委員: |
李文猶
Wen-Yo Lee 黃志良 Chih-Lyang Hwang 何昭慶 Chao-Ching Ho 施慶隆 Ching-Long Shih |
學位類別: |
碩士 Master |
系所名稱: |
電資學院 - 電機工程系 Department of Electrical Engineering |
論文出版年: | 2020 |
畢業學年度: | 108 |
語文別: | 中文 |
論文頁數: | 69 |
中文關鍵詞: | 深度學習 、撞球策略 、Yolo v3 、Mask R-CNN |
外文關鍵詞: | deep learning, billiard strategy, Yolo v3, Mask R-CNN |
相關次數: | 點閱:393 下載:8 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本論文旨在應用深度學習辨識輸入影像中的彩色球,並提出撞球策略以比
較不同目標球的入袋難易度。本文將一網路攝影機安裝於機械臂終端,並在攝
影機下方安裝一氣壓缸作為撞球桿,使用影像回受控制校正目標球、母球以及
球桿終點之直角坐標位置。本文首先比較以同樣的訓練資料所訓練出來的兩個
目標辨識模型Yolo v3 以及Mask R-CNN 在撞球影像辨識上的成效。因Mask RCNN
之召回率較高且影像座標位置的誤差較小,故使用Mask R-CNN 檢測輸入
影像中的彩色球。為得到入袋難度最低的目標球,本文提出撞球策略用以分析
所有子球之直射球或組合球的入袋難易度。最後將1 或2 顆子球隨機擺放在撞
球桌上,使用直射球或組合球將目標球撞擊入袋,重複執行155 次可發現入袋
成功時的評估函數通常大於入袋失敗時的評估函數,因此本文提出之撞球策略
用於判斷入袋難易度是十分可靠的。
The objective of this study is to detect color balls in an input image using deep
learning, and to propose a billiard strategy to compare the difficulty of pocketing
different target balls. A webcam is equipped at the end of a manipulator. A pneumatic
cylinder is equipped under the webcam as a cue stick. To align the position of the cue
ball, the target ball, and the end of the cue stick in the cartesian coordinate, a
manipulator is commanded to perform visual alignment. In this study we compare the
performance of two object detection models, Yolo v3 and Mask R-CNN, trained by
the same training data. Since Mask R-CNN has higher Recall and the lower error in
the image position, Mask R-CNN is used to detect color balls in an input image. In
order to get the target ball with the lowest difficulty in pocketing, this study propose a
billiard strategy to analyze the difficulty of using direct or combination shot to pocket
different target balls. Pocketing 1 or 2 color balls randomly placed on the billiard table
and use a direct or combination shot to hit the target ball into the pocket. After
repeating the process 155 times, it can be found that the successful evaluation
functions are usually higher than the failed evaluation functions. Therefore, the
billiard strategy proposed in this paper is very reliable to compare the difficulty of
pocketing different target balls.
[1] Wei Shen and Lifang Wu, “A method of billiard objects detection based on
Snooker game video,” 2010 2nd International Conference on Future Computer
and Communication. IEEE, 2010.
[2] Yan Ling, Shaobin Li, Pin Xu and Bin Zhou, “The detection of multi-objective
billiards in snooker game video,” 2012 Third International Conference on
Intelligent Control and Information Processing. IEEE, 2012.
[3] Jungkil Park and Jaebyung Park, “CAMshift Algorithm based on geometric
information for fast billiard ball tracking,” 2015 12th International Conference
on Ubiquitous Robots and Ambient Intelligence (URAI). IEEE, 2015.
[4] Shraddha Mane and Supriya Mangale, “Moving object detection and tracking
using convolutional neural networks,” 2018 Second International Conference on
Intelligent Computing and Control Systems (ICICCS). IEEE, 2018.
[5] Matija Buric, Miran Pobar and Marina Ivasic-Kos, “Ball detection using Yolo
and Mask R-CNN,” 2018 International Conference on Computational Science
and Computational Intelligence. IEEE, 2018.
[6] Z.M.Lin, J.S.Yang and C.Y.Yang, “Grey decision-making for a billiard robot,”
2004 IEEE International Conference on Systems, Man and Cybernetics (IEEE
Cat. No.04CH37583). IEEE, 2004.
[7] Michael Smith, “Pick pocket: a computer billiards shark,” Artificial Intelligence
171(16-17):1069-1091, 2007.
[8] Jean-Francois Landry and Jean Pierre Dussault, “AI optimization of a billiard
player,” Journal of Intelligent and Robotic Systems, 2007.
[9] Jr-Syu Yang and You-Mu Chen, “The design of jump shot decision-making
56
system for a billiard robot,” 2010 IEEE Conference on Robotics, Automation
and Mechatronics. IEEE, 2010.
[10] Chihhsiong Shih, “Analyzing and comparing shot planning strategies and their
effects on the performance of an augment reality based billiard training system,”
international journal of information technology & decision making Vol. 13, 2014.
[11] Tatsuyuki Nezu, “Grey system research and application in the pool game,”
International Conference on Electromechanical Control Technology and
Transportation. ICECTT, 2015.
[12] Jiaying Gao, Ming Zhu, Haoquan Liang and Xiao Guo, “Design of the multiple
neural network compensator for a billiard robot,” 2015 IEEE 12th International
Conference on Networking, Sensing and Control. IEEE, 2015.
[13] Jiaying Gao, Qiuyang He and Zhixin Zhan, “Dynamic modeling based on fuzzy
neural network for a billiard robot,” 2016 IEEE 13th International Conference on
Networking, Sensing, and Control (ICNSC). IEEE, 2016.
[14] Yu Chen and Yujun Li, “Macro and micro reinforcement learning for playing
nine-ball pool,” 2019 IEEE Conference on Games (CoG). IEEE, 2019.
[25] Kuei Gu Tung, Sheng Wen Wang, Wen Kai Tai, Der Lor Way and Chin Chen
Chang, “Toward human-like billiard AI bot based on backward induction and
machine learning,” 2019 IEEE Symposium Series on Computational
Intelligence (SSCI). IEEE, 2019.
[16] Bo-Ru Cheng, Je-Ting Li and Jr-Syu Yang, “Design of the neural-fuzzy
compensator for a billiard robot,” IEEE International Conference on Networking,
Sensing and Control, 2004. IEEE, 2004.
[17] Jie Shi, Yali Zhou and WeiJie Xia Qizhi Zhang, “Target detection based on
improved Mask R-CNN in service robot,” 2019 Chinese Control Conference
(CCC). IEEE, 2019.
57
[18] Thomas Nierhoff, Omiros Kourakos and Sandra Hirche, “Playing pool with a
dual-armed robot,” 2011 IEEE International Conference on Robotics and
Automation. IEEE, 2011.
[19] Daiki Suzuki and Kouhei Ohnishi, “A skill proficiency evaluation method of
individual based on accuracy and repeatability,” IECON 2012 - 38th Annual
Conference on IEEE Industrial Electronics Society. IEEE, 2012.
[20] Chihhsiong Shih and William C. Chu, “A vision based interactive billiard ball
entertainment system,” 2007 First IEEE international workshop on digital game
and intelligent toy enhanced learning (DIGITEL'07). IEEE ,2007.
[21] Chen-Wei Chou, Ming-Chun Tien and Ja-Ling Wu, “Billiards wizard: A tutoring
system for broadcasting nine-ball billiards videos,” 2009 IEEE International
Conference on acoustics, speech and signal processing. IEEE, 2009.
[22] Masashi Mishima and Akira Suganuma, “Development of stance correction
system for billiard beginner player,” 2016 International Symposium on
Intelligent Signal Processing and Communication Systems (ISPACS). IEEE,
2016.
[23] Chun-Chieh Hsu, Hou-Chun Tsai, Hua-Tsung Chen, Wen-Jiin Tsai and Suh-Yin
Lee, “Computer-assisted billiard self-training using intelligent glasses,” 2017
14th International Symposium on Pervasive Systems, Algorithms and Networks
& 2017 11th International Conference on frontier of Computer Science and
technology & 2017 third International Symposium of Creative Computing
(ISPAN-FCST-ISCC). IEEE, 2017.
[24] Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun, “Deep residual
learning for image recognition,” 2016 IEEE Conference on Computer Vision and
Pattern Recognition (CVPR). IEEE, 2016.
[25] Tsung-Yi Lin, Piotr Dollár, Ross Girshick, Kaiming He, Bharath Hariharan and
58
Serge Belongie, “Feature pyramid networks for object detection,” 2017 IEEE
Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2017.