臺灣博碩士論文加值系統

交通工具的普及雖然為人們帶來了許多便利，同時也衍生出不少的問題。根據
2017 年台灣地區年交通部機動車輛登記數統計，機動車數量、道路交通事故肇事
數量及道路交通違規數量有的增加有的減少，但是這些為數不少的事故及違規事
件大部分還是依賴人工進行處理。為求方便，這些問題若能搭配一套完善的系統
協助處理，減少大量的人力消耗在處理這些問題上，便能將人員做更好的分配及
安排。
因此，我們於本論文中提出一種車牌辨識的演算法。對於取得的影像先進行車
牌的擷取，讓影像呈現水平狀態，再將影像做轉換並去除雜訊，接著將字元萃取
出來與資料庫做比對，最後輸出比對的結果。
而我們皆使用自網路取得的車牌影像來作測試。除了有標準的車牌影像之外，
也有各式各樣不同拍攝角度的車牌影像，只要是做跟車牌影像辨識有關研究的人
員，都可以到這些網站尋找適合測試的影像。相較於自行拍攝的影像，更具有參
考的價值。
從測試完的結果可以得知，對於大部分的車牌影像都可以辨識出結果，但若有
妨礙字元辨識的部分，例如車牌本身的髒汙及拍攝時產生的雜訊等等，或是車牌
傾斜角度過大的情形都是影響辨識的因素。

Although the popularity of transportation brings great convenience to people, it
still causes a lot of problems. According to the statistics of motor vehicle registration
from the Ministry of Transportation and Communications in Taiwan in 2017, there
are increases and decreases in the number of motor vehicles, road traffic accidents,
and road traffic offences. However, most of these traffic accidents and traffic offences
are handled manually. For the sake of convenience, if these problems can be dealt
with a complete system, a large amount of manpower in these cases can be reduced,
personnel distributions and arrangements can be better.
Therefore, in this essay, we propose an algorithm for recognition of license plates.
First, we take out the part of the license plate in the images and we check whether
the license plate is presented in a horizontal state or not. Next, we convert the
images and remove the noise. Then we extract the characters and compare them
with those in the database. At last, we obtain the results of comparison.
For the images of the license plates that we use for testing, they were taken from
open data on the Internet. In addition to standard images, there are images taken
from different angles. People who do the research on recognition of license plates
can find images that are suitable for testing on these websites. Compare with photos
taken by ourselves, these images are more referential.
Based on the results of the test, we have known that most of the images of the
license plates can be identified. However, if the characters are obstructed, such as
the blurs of the license plates and the noise generated when photographing etc., or
the license plates incline too much, these factors will affect the process of identification.

第 1 章 導論 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 研究背景. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 文獻回顧. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 研究動機. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 文章架構. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
第 2 章 車牌的擷取及處理. . . . . . . . . . . . . . . . . . . . . . . .4
2.1 邊緣偵測. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 直線偵測. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 仿射轉換. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 彩色圖轉灰階圖. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5 除雜訊. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
2.6 數學形態學轉換. . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
第 3 章 車牌字元的辨識. . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1 字元萃取. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 字元分割. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3 比對資料庫. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
3.4 演算法. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
第 4 章 車牌辨識測試 . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1 測試環境. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
4.2 測試成功的結果. . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
4.3 測試失敗的結果. . . . . . . . . . . . . . . . . . . . . . . . . . . . .26
第 5 章 結論與未來展望 . . . . . . . . . . . . . . . . . . . . . . . . . .30
5.1 結論. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2 未來展望. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30
參考文獻. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
附錄. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35

[1] E. Arias-Castro, and D. L. Donoho (2009), ”Does median filtering truly preserve
edges better than linear filtering？”, The Annals of Statistics, vol. 37, no. 3,
1172-1206.
[2] K. Arakawa (1996), ”Median filter based on fuzzy rules and its application to
image restoration”, Fuzzy Sets and Systems, vol. 77, no. 1, 3-13.
[3] R. Van Den Boomgaard, and R. Van Balen (1992), ”Methods for fast morphological
image transforms using bitmapped images”, CVGIP: Graphical Models
and Image Processing, vol. 54, no. 3, 252-258.
[4] J. F. Canny, ”Finding edges and lines in images”, M.I.T. Computer Science and
Artificial Intelligence Laboratory, Boston, Cambridge, 1983.
[5] J. Canny (1986), ”A computational approach to edge detection”, IEEE Transactions
on Pattern Analysis and Machine Intelligence, vol. PAMI-8, no. 6, 679-
698.
[6] R. O. Duda, P. E. Hart (1972), ”Use of the hough transform to detect lines and
curves in pictures”, Communications of Association for Computing Machinery,
vol. 15, no. 1, 11-15.
[7] P. V. C. Hough (1962), ”Method and means for recognizing complex patterns”,
U.S. Patent No. 3,069,654, Washington, DC: U.S. Patent and Trademark Office.
[8] J. Illingworth, J. Kittler (1988), ”A survey of the Hough transform”, Computer
Vision, Graphics, and Image Processing, vol. 44, no. 1, 87-116.
[9] M. I. Khalil (2010), ”Car plate recognition using the template matching
method”, International Journal of Computer Theory and Engineering, vol. 2,
no. 5, 683-687.
[10] J. S. Lim, ”Two-dimensional signal and image processing”, Prentice Hall,
United States, New Jersey, 1990.
[11] G. Matheron, and J. Serra (2002), ”The birth of mathematical morphology”,
In Proceedings of the 6th International Symposium Mathematical Morphology,
1 April, Sydney, Australia, 1-16.
[12] S. G. Patel (2013), ”Vehicle license plate recognition using morphology and
neural network”, International Journal on Cybernetics and Informatics, vol. 2,
no. 1, 1-7.
[13] W. Rong, Z.Li, W.Zhang, and L.Sun (2014), ”An improved Canny edge detection
algorithm”, In Proceeding of 2014 Electrical and Electronics Engineers
International Conference on Mechatronics and Automation, 3-6 August, Tianjin,
China, 577-582.
[14] R. Sharma (2015), ”Automatic license plate based smart vehicle validation and
security by gate control and e-mail send”, International Journal of Computer
Science and Information Technology, vol. 6, no. 2, 952-957.
[15] S. Ozbay, E. Ercelebi (2005), ”Automatic vehicle identification by plate recognition”,
World Academy of Science, Engineering and Technology, International
Journal of Electrical, Computer, Energetic, Electronic and Communication Engineering,
vol. 1, no. 9, 1410-1413.
[16] J. Serra, ”Image analysis and mathematical morphology”, Academic Press,
United Kingdom, London, 1982.
[17] P. Soille, ”Morphological image analysis:principles and applications”, Springer
Science & Business Media, Germany, Heidelberg, 1999.
[18] B. Tiwari, A. Sharma, M. G. Singh, B. Rathi (2016), ”Automatic vehicle number
plate recognition system using MATLAB”, International Organization of
Scientific Research Journal of Electronics and Communication Engineering, vol.
11, no. 4, 10-16.
[19] License Plates of the World。取自http://www.worldlicenseplates.com
[20] License Plate Mania - Taiwan。取自http://www.licenseplatemania.com/
landenpaginas/taiwan.htm
[21] Taiwanease, making taiwan easy！ 取自https://taiwanease.com/en/
forums/new-taiwan-license-plates-coming-on-dec-17-t9695.html?
nomobile=1
[22] 交通部統計查詢網。機動車輛數常用查詢網頁。取自http://stat.motc.
gov.tw/mocdb/stmain.jsp?sys=100&funid=a3301
[23] 中華民國交通部公路總局。監理服務網。取自https://www.mvdis.gov.tw/
m3-emv-plate/webpickno/member/queryFitNewPlate#
[24] 中華民國交通部公路總局。號牌型式及編碼規則表。取自https://www.thb.
gov.tw/page?node=92d4a6e2-9afb-464d-a2eb-2be8d26d8d89