車牌辨識應用深度學習於Android 行動裝置

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：53

、訪客IP：18.219.81.129

姓名

廖首名(Shou-Ming Liao) 查詢紙本館藏

畢業系所

電機工程學系

論文名稱

車牌辨識應用深度學習於Android 行動裝置
(Application of deep learning in license plate recognition to Android mobile devices)

相關論文

★ 獨立成份分析法於真實環境中聲音訊號分離之探討	★ 口腔核磁共振影像的分割與三維灰階值內插
★ 數位式氣喘尖峰氣流量監測系統設計	★ 結合人工電子耳與助聽器對中文語音辨識率的影響
★ 人工電子耳進階結合編碼策略的中文語音辨識成效模擬--結合助聽器之分析	★ 中文發聲之神經關聯性的腦功能磁振造影研究
★ 利用有限元素法建構3維的舌頭力學模型	★ 以磁振造影為基礎的立體舌頭圖譜之建構
★ 腎小管之草酸鈣濃度變化與草酸鈣結石關係之模擬研究	★ 口腔磁振影像舌頭構造之自動分割
★ 微波輸出窗電性匹配之研究	★ 以軟體為基準的助聽器模擬平台之發展-噪音消除
★ 以軟體為基準的助聽器模擬平台之發展-回饋音消除	★ 模擬人工電子耳頻道數、刺激速率與雙耳聽對噪音環境下中文語音辨識率之影響
★ 用類神經網路研究中文語音聲調產生之神經關聯性	★ 教學用電腦模擬生理系統之建構

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2026-9-1以後開放)

摘要(中)

此次的研究是以深度學習神經網路來進行車牌定位與字元辨識，本研究主要是使用YOLOv7 (You Only Look Once v7)進行訓練並辨識目標物件，相較之前的最突出的實時物件偵測技術，像是YOLOv5 (You Only Look Once v5)與高效目標檢測模型 (EfficientDet)等物件偵測模型，降低了40% 參數量、50%運算量，並有更快的速度與正確率，因此決定採用YOLOv7。鑒於傳統車牌辨識方法需要經過二值化與侵蝕篩選等等，或辨識環境受限固定角度、光源與位置，所以本研究目標是利用超解析模型、低光源演算法搭配 YOLOv7，完成多物件辨識且不受限多種環境進行快速辨識。
本次實驗包含各式車輛且遠近不同的街景圖，總計車牌訓練資料 5524張，字元訓練資料 4676 張，測試資料300張包含各式複雜環境(遠景、歪斜、模糊、昏暗等等)，車牌定位正確率可以達到98.5%，三公尺內合理拍攝則超過99%，且召回率為98.1%，F1-Score為98.3%。而字元辨識正確率達99.3%，召回率為98.6%，F1-Score為98.95%。其中因YOLOv7的深度學習框架使用Pytorch，但因為未來需要用在行動裝置，因此需要更輕量化以減小中央處理器 (Central Processing Unit, CPU)負擔，因此需要經過多種深度學習框架的轉換，從開放式交換神經網路 (Open Neural Network Exchange, ONNX)、張量流(Tensorflow, TF)、輕量化張量流 (Tensorflow Lite, TFLite)的轉換。最後利用Andorid Studio實作 APP 程式，並於 Android 行動工業電腦上完成三公尺內合理拍攝車牌的99%正確率與即時偵測，以便未來在收費員與警察查緝之利用。

摘要(英)

This study focuses on utilizing deep learning neural networks to perform license plate detection and character recognition. The primary approach employed in this research is the training and utilization of YOLOv7 (You Only Look Once V7) for object detection. Compared to previous prominent real-time object detection techniques like YOLOv5 and EfficientDet, YOLOv7 offers a remarkable reduction of 40% in parameter count and 50% in computational load while maintaining faster inference speeds and improved precision.
Consequently, YOLOv7 was chosen as the preferred model for this study.Given that conventional license plate recognition methods often involve preprocessing steps such as binarization, erosion filtering, and are limited by fixed angles, lighting conditions, and positions, this research aims to address these limitations by leveraging super-resolution models and low-light algorithms in conjunction with YOLOv7. The objective is to achieve fast and robust license plate recognition across diverse environmental conditions. The experimental dataset comprises a variety of street scenes featuring different vehicles at varying distances. In total, there are 5,524 images for license plate training and 4,676 images for character training. The test dataset consists of 300 images that encompass various complex environments such as distant views, tilting, blurriness, dim lighting, and more. The accuracy of license plate localization reaches 98.5%, while for reasonably captured plates within a three-meter range, it exceeds 99%. The character recognition accuracy is 99.3%. Additionally, the recall for license plate localization is 98.1%, and the F1-Score is 98.3%. For character recognition, the recall is 98.6%, and the F1-Score is 98.95%.
As YOLOv7 utilizes the PyTorch deep learning framework, efforts were made to ensure compatibility with mobile devices, necessitating model lightweighting to alleviate the computational burden on Central Processing Unit (CPU). Consequently, the models underwent conversion across multiple deep learning frameworks, including Open Neural Network Exchange (ONNX), TensorFlow (TF), and TensorFlow Lite (TFLite).Finally, an Android Studio application was developed to deploy the system, achieving a 99% accuracy in license plate recognition for license plates reasonably captured within a three-meter range, and enabling real-time detection on an Android mobile industrial computer.

關鍵字(中)

★ 深度學習
★ 車牌辨識

關鍵字(英)

★ Android mobile device
★ deep learning

論文目次

目錄:
摘要 I
Abstract III
目錄 IV
圖目錄 VIII
表目錄 XIII
第一章緒論 1
1.1研究動機 1
1.2基本車牌樣式與車牌字元介紹 3
1.2.1車牌樣式 3
1.2.2車牌字元 4
1.3文獻探討 4
1.3.1傳統車牌辨識演算法 4
1.3.2深度學習神經網路模型 6
1.4論文架構 10
第二章深度學習神經網路簡述 11
2.1類神經網路概述 11
2.1.1感知器 (Perceptron) 11
2.1.2激活函數 (Activation Function) 12
2.1.3損失函數 (Loss Function) 15
2.1.4 卷積神經網路 (Convolutional Neural Network, CNN) 17
2-2 物件偵測技術 21
2.2.1 YOLO家族的發展(v1到v3) 21
2.2.2 YOLO使用到的技術 23
2.2.3 YOLOv4 26
2.2.4 YOLOv7 33
第三章相關演算法架構與移動裝置APP優化方法 45
3.1資料收集 46
3.1.1 RS35實拍 46
3.1.2 Iphone實拍 47
3.1.3 RoboFlow 47
3.1.4 合作公司提供資料 48
3.1.5字元資料(由所有資料集中進行截取與切割) 48
3.1.6資料標記 49
3.2 超解析影像重建 50
3.2.1 超解析卷積神經網路 (Super-Resolution Convolutional Neural Network, SRCNN) 50
3.3 低光源影像轉換 51
3.3.1 暗通道先驗 (Dark Channel Prior) 51
3.4 深度學習框架 53
3.4.1 蟒蛇火炬 (Pytorch) 54
3.4.2 開放式交換神經網路 (Open Neural Network Exchange, ONNX) 54
3.4.3 張量流 (Tensorflow)與輕量化張量流 (Tensorflow Lite) 55
3.4.4 Tensorflow Lite架構 56
3.5 模型與APP優化 57
3.5.1 YOLOv7-Tiny 58
3.5.2 模型量化 (Quantization) 58
3.5.3 模型剪枝 (Pruning) 60
3.5.4 模型輸入影像大小調整 (Input Size) 61
3.5.5 CPU單線程與多線程 61
第四章研究結果與討論 63
4.1 實驗設備 63
4.2 測試圖片 65
4.3 軟體端影像辨識結果 65
4.3.1 YOLOv7車牌定位結果 67
4.3.2 YOLOv7字元辨識結果 72
4.3.3 超解析模型影像重建結果 75
4.3.4 低光源影像轉換結果 77
4.3 Android行動裝置與APP之影像辨識結果 79
4.3.1 APP版面配置 79
4.3.2 APP辨識結果 80
4.4 結果與討論 85
4.4.1 評估指標Precision與 Recall的差別 85
4.4.2準確率與辨識速度 86
第五章結論與未來展望 87
5.1 結論 87
5.2 未來展望 89
參考文獻 91

參考文獻

Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.(2017). "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications, " arXiv:1704.04861

Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton.(2017). "ImageNet classification with deep convolutional neural networks, " Communications of the ACM, Volume 60, Issue 6, June 2017, pp. 84–90.

Alexey Bochkovskiy,Hong-Yuan Mark Liao,Chien-Yao Wang.(2020).
"YOLOv4: Optimal Speed and Accuracy of Object Detection, "
arXiv:2004.10934

Chao-Ho Chen, Tsong-Yi Chen, Min-Tsung Wu, Tsann-Tay Tang, Wu-Chih Hu. "License Plate Recognition for Moving Vehicles Using a Moving Camera, "2013 Ninth International Conference on Intelligent Information Hiding and Multimedia Signal Processing, 2013, pp. 497-500.

Chao Dong, Chen Change Loy, Kaiming He, Xiaoou Tang.(2014). "Image Super-Resolution Using Deep Convolutional Networks, "arXiv:1501.00092

Christian Ledig, Lucas Theis, Ferenc Huszar, Jose Caballero, Andrew
Jiwon Kim, Jung Kwon Lee, Kyoung Mu Lee.(2015). "Accurate Image Super-Resolution Using Very Deep Convolutional Networks," arXiv:1511.04587

Cunningham, Alejandro Acosta, Andrew Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, Wenzhe Shi.(2017). "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network, "arXiv:1609.04802

C. Lin, Y. Lin, and W. Liu, "An efficient license plate recognition system using convolution neural networks, " in 2018 IEEE International Conference on Applied System Invention (ICASI), 2018, pp. 224–227.
Chien-Yao Wang, Alexey Bochkovskiy, Hong-Yuan Mark Liao.(2022). "YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors, "arXiv:2207.02696

E. Ullah, R. Nawaz, J. Iqbal.(2013). "Single image haze removal using improved dark channel prior, "2013 5th International Conference on Modelling, Identification and Control (ICMIC), pp. 245-248.

G. -W. Chen, C. -M. Yang and T. -U. İk, "Real-Time License Plate Recognition and Vehicle Tracking System Based on Deep Learning," 2021 22nd Asia-Pacific Network Operations and Management Symposium (APNOMS), 2021, pp. 378-381.

I. H. El-Shal, O. M. Fahmy and M. A. Elattar, "License Plate Image Analysis Empowered by Generative Adversarial Neural Networks (GANs), " in IEEE Access, vol. 10, pp. 30846-30857, 2022

Joseph Redmon, Ali Farhadi.(2018). "YOLOv3: An Incremental Improvement, " arXiv:1804.02767

J.L. LAI, "High efficient license plate recognition system based on deep learning" Master’s thesis, National Yunlin University of Science and Technology, 2019.

J. -Y. Sung, S. -B. Yu and S. -h. P. Korea, "Real-time Automatic License Plate Recognition System using YOLOv4," 2020 IEEE International Conference on Consumer Electronics - Asia (ICCE-Asia), 2020, pp. 1-3.

Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.(2015). "Deep Residual Learning for Image Recognition, " arXiv:1512.03385

Li Shuangfeng(Google TensorFlow Team, Beijing 100190).(2020). "TensorFlow Lite：on—DeVice Machine Learning Framework, " Journal of Computer Research and Development, pp. 1839-1853.

N. Kanopoulos, N. Vasanthavada and R.L. Baker.(1988). "Design of an image edge detection filter using the Sobel operator, "1988 IEEE Journal of Solid-State Circuits, pp. 358-367.

Yusef Alborzi, Talayeh Sarraf Mehraban, Javad Khoramdel, Ali Najafi Ardekany. "Robust Real time Lightweight Automatic License plate Recognition System for Iranian License Plates, " 2019 7th International Conference on Robotics and Mechatronics (ICRoM), 2019, pp. 352-356.

Z. T. Soghadi and C. Y. Suen, "License plate detection and recognition by convolutional neural networks, " in Pattern Recognition and Artificial Intelligence, Y. Lu, N. Vincent, P. C. Yuen, W.S. Zheng, F. Cheriet, and C. Y. Suen, Eds. Cham: Springer International Publishing, 2020, pp. 380–393.

許博翔,林政宏"A lightweight Detection Model for License plate Recognition, "台灣碩博士論文 110年9月

林立,陳慶瀚.(2010). "A Based on Dual-Core Processor License Plate Recognition System, "中央大學碩博士論文民國99年6月

陳翔傑,王文俊.(2005). "自動化車牌辨識系統設計, " 中央大學碩博士論文民國94年6月

物件偵測NMS (2018)。2023 年 6 月 26 日取自
https://chih-sheng-huang821.medium.com/%E6%A9%9F%E5%99%A8-%E6%B7%B1%E5%BA%A6%E5%AD%B8%E7%BF%92-%E7%89%A9%E4%BB%B6%E5%81%B5%E6%B8%AC-non-maximum-suppression-nms-aa70c45adffa

公路總局之車種與車牌類別(2023)。2023年6月27取自https://www.thb.gov.tw/cp.aspx?n=102

李謦伊, "YOLO演進—3—YOLOv4詳細介紹". available at https://medium.com/ching-i/yolo%E6%BC%94%E9%80%B2-3-yolov4%E8%A9%B3%E7%B4%B0%E4%BB%8B%E7%B4%B9-5ab2490754ef since 2020. 2023/06/27

王進德.(2006).類神經網路與模糊控制理論入門與應用.全華圖書股份有限公司.臺北市.

指導教授

吳炤民(Chao-Min Wu)

審核日期

2023-8-8

推文