跳到主要內容

臺灣博碩士論文加值系統

(44.200.140.218) 您好!臺灣時間:2024/07/18 05:04
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:鄭益峻
研究生(外文):CHENG, YI-CHUN
論文名稱:利用半監督式學習與資料增強提升YOLO車牌偵測之性能
論文名稱(外文):Improve the performance of YOLO-based license plate detection with semi-supervised learning and data augmentation
指導教授:陳慶逸陳慶逸引用關係
指導教授(外文):CHEN, CHING-YI
口試委員:陳慶瀚李棟良
口試委員(外文):CHEN, CHING-HANLI, TUNG-LIANG
口試日期:2021-01-22
學位類別:碩士
校院名稱:銘傳大學
系所名稱:電腦與通訊工程學系碩士班
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2021
畢業學年度:109
語文別:中文
論文頁數:52
中文關鍵詞:車牌偵測YOLO車牌偵測半監督式學習資料擴增
外文關鍵詞:License plate detectionYOLO license plate detectionSemi-supervised learningData augmentation
Facebook:https://www.facebook.com/aron3333/
相關次數:
  • 被引用被引用:1
  • 點閱點閱:485
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
要實現路邊車輛之自動車牌偵測工作,無論是從極高視角或是側邊拍攝的取景角度,所得到的車輛影像都有極端斜角的問題,同時也可能發生影像中景物複雜的情況,造成在影像分析上的困難。傳統的物件偵測演算法往往以特徵擷取和分類器二階段的整合來實現,到了深度學習的時代,使用卷積神經網路能夠擷取到影像中更豐富、更深層次的特徵,因此相關的作法也成為了物件偵測的主流。然而,深度學習方法需要大量的加標資料才能有效訓練模型;但對多數的影像應用而言,加標資料並不易取得。本研究提出一個可適用在有限加標資料集的Yolo車牌偵測系統;該方法使用半監督式學習的方式產生大量的加標影像,最後再整合灰階化與小波轉換等影像處理的作法進行資料擴增,即使加標的訓練資料有限,經由本研究所述的方法仍可有效提升 Yolo 車牌偵測模型的偵測成功率。
To realize the automatic license plate detection of roadside vehicles, whether it is from a very high viewing angle or a viewing angle shot from the side, the obtained vehicle image has the problem of extreme oblique angles, and the scene in the image may also be complicated. , Causing difficulties in image analysis. Traditional object detection algorithms are often implemented by the two-stage integration of feature extraction and classifiers. In the era of deep learning, the use of convolutional neural networks can capture richer and deeper features in the image, so relevant The approach has also become the mainstream of object detection. However, the deep learning method requires a large amount of spiked data to effectively train the model; however, for most imaging applications, the spiked data is not easy to obtain. This research proposes a Yolo license plate detection system that can be applied to a limited set of labeled data; this method uses semi-supervised learning to generate a large number of labeled images, and finally integrates image processing methods such as grayscale and wavelet transformation. Data amplification, even if the training data to be added is limited, the method described in this study can still effectively improve the detection success rate of the Yolo license plate detection model.
摘要 i
Abstract ii
目錄 iii
圖目錄 v
表目錄 vii
第一章 緒論 1
1.1研究背景與動機 1
1.2文獻探討 2
第二章 相關研究 4
2.1 影像前處理 4
2.1.1 彩色影像灰階化 4
2.1.2 小波轉換 5
2.1.3 離散小波轉換(DWT) 8
2.1.4 對稱式遮罩離散小波轉換(SMDWT) 10
2.2 卷積神經網路 11
2.2.1 卷積層 12
2.2.2 Subsampling/ Pooling Layer(降取樣/池化層) 14
2.2.3 激活函數 15
2.2.4 全連接層 18
2.3 YOLO演算法 18
2.4 半監督式學習 21
第三章 基於YOLO v3模型之車牌偵測系統 25
3.1 資料擴增 (Data augmentation) 28
3.2 半監督式學習(semi-supervised learning) 30
第四章 實驗結果 31
4.1 實驗環境 31
4.2 車牌影像資料資料集 32
4.3 不同影像處理之模型實驗 33
4.4 半監督式學習模型實驗 35
4.5 新加標影像與資料擴增實驗 36
4.6 結果討論 38
第五章 結論 41
參考文獻 42
[1] J. Hsieh, S. Yu, and Y. Chen, “Morphology-based license plate detection from complex scenes,” IEEE International Conference on Pattern Recognition, 2002: pp. 176 – 179.
[2] J. Guo, and Y. Liu, “License plate localization and character segmentation with feedback self-learning and hybrid binarization techniques,” IEEE Transactions on Vehicular Technology, 2008, 57(3): pp. 1417 – 1424.
[3] 王子豪,基於SIFT演算法進行車牌認證,國立中央大學資工系碩士論文, 2017年
[4] X. He, H. Zhang, W. Jia, Q. Wu, and T.Hintz, “Combining global and local features for detection of license plates in a video,” Proceedings of Image and Vision Computing, New Zealand, 2007: pp. 288–293.
[5] W. Ho, H. Lim, and Y. Tay, “Two-stage License Plate Detection using Gentle Adaboost and SIFT-SVM,” 2009 First Asian Conference on Intelligent Information and Database Systems, 2009.
[6] Y. Zhao, J. Gu, C. Liu, and S. Han, “License Plate Location Based on Haar-Like Cascade Classifiers and Edges,” 2010 Second WRI Global Congress on Intelligent Systems, 2010.
[7] S. Wang, and H. Lee, “A cascade framework for a real-time statistical plate recognition system,” IEEE Transactions on Information Forensics and Security, 2007, 2(2): pp. 267–282.
[8] A. Krizhevsky, I. Sutskever, and G. Hinton, “Imagenet classification with deep convolutional neural networks,” in Advances in neural information processing systems, pp. 1097–1105, 2012.
[9] J. Redmon, and A. Farhadi, “Yolov3: An incremental improvement,” arXiv: 1804.02767, 2018.
[10] J.-L. Wu, C.-H. Chen and C.-F. Chen, “Numerical inversion of Laplace transform using Haar wavelet operational matrices,” IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, Vol.48, pp.120-122,2001.
[11] C.-H. Hsia, J.-M. Guo, and J.-S. Chiang, “Improved low complexity algorithm for 2-D integer lifting-based discrete wavelet transform using symmetric mask-based scheme,” IEEE Transactions on Circuits and Systems for Video Technology, Vol.19, No.8, pp.1202-1208, 2009.
[12] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-Based Learning Applied to Document Recognition,” in Proceedings of the IEEE, Vol.86, Issue 11, Nov 1998.
[13] 陳彥志,使用新式特徵參數於車牌辨識系統效能改善之研究,國立勤益科技大學電機工程學系碩士論文,2013年6月
[14] 杜柏宏,訓練資料集自我擴增問題之分析-以YOLO分類器為例,國立中山大學電機工程學系碩士論文,2019年3月
[15] D. Rumelhart, G. Hinton and R. Williams, “Learning representations by back-propagating errors, ” Nature,323:533-536,1986
[16] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi. “You only look once: Unified,real-time object detection,” arXiv preprint arXiv: 1506.02640,2015.
[17] O. Chapelle, B. Scholkopf, and A. Zien. “Semi-supervised learning (chapelle, o. et al., eds.; 2006)[book reviews].” IEEE Transactions on Neural Networks, 20(3):542–542, 2009.
[18] M. Mulder, D. Abbink, M. Paassen, and M. Mulder, “Design of a haptic gas pedal for active car-following support,” IEEE Trans. Intell. Transp. Syst., vol.12, no. 1, pp. 268-279, Mar. 2011.
[19] C.-N. Anagnostopoulos, I. Anagnostopoulos, I. Psoroulas, V. Loumos, and E. Kayafas, “License plate recognition from still images and video sequences: A survey,” IEEE Trans. Intell. Transp. Syst., vol. 9, no. 3, pp. 377-391, Sep. 2008.
[20] D. Zheng , Y. Zhao , and J. Wang, “An Efficient Method of License Plate Location,” in 2009 First International Conference on Information Science and Engineering, Nov. 2005.
[21] 簡維皇,用於路口查驗贓車的車牌辨識系統,國立中央大學資訊工程學系碩士論文,2011年6月
[22] 陳昱丞,基於卷積神經網路之車牌辨識系統,國立交通大學電控工程研究所碩士論文,2017年7月
[23] 許亨安,基於特徵值與類神經網路之車牌辨識,國立高雄應用科技大學電機工程學系碩士論文,2011年7月
[24] 林永鑫,植基於卷積神經網路之高效能車牌辨識系統,國立臺灣師範大學電機工程學系碩士論文,2018年8月
[25] 曾士豪,自動化車牌辨識系統,國立交通大學電控工程研究所碩士論文,2016年6月
[26] 楊俊元,適用於低解析度影像處理方法之演化式提升離散小波轉換濾波器設計,銘傳大學電腦與通訊工程學系碩士班碩士論文,2015年

電子全文 電子全文(網際網路公開日期:20250219)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊