Vehicle license plate location using active learning AdaBoost algorithm and color feature
Article Text (Baidu Translation)
-
摘要: 人工选取少量的车牌区域和非车牌区域, 采用积分图法快速提取Haar-like扩展特征, 构成初始训练样本。使用AdaBoost算法训练样本产生一个初始分类器, 经过主动学习过程, 产生一个用于车牌检测的强分类器。利用Cascade结构检测法进行车牌的粗定位, 通过提取边缘颜色对, 对候选区域进行验证, 实现车牌区域的精确定位。对不同光照条件及车牌污损等复杂情况下的车牌图像进行了定位测试。测试结果表明: 车牌的粗定位率和精确定位率分别为98.3%、97.1%, 平均定位时间小于0.1 s, 因此, 该方法有较好的车牌定位效果和定位准确率。Abstract: A small amount of license plate areas and non-license plate areas were selected, and Haar-like extended features were extracted by using the integration diagram method to obtain initial training samples. An initial classifier was generated by training the samples with AdaBoost algorithm. A strong classifier for license plate detection was obtained in the active learning procedure. The coarse location of license plate was implemented by using the cascade structure detection method. The candidate region was verified to get the precise location of license plate area by extracting edge-color pairs. The method was applied into the test of vehicle license plate location under different illumination and defaced circumstances. Test result indicates that the coarse location rate of license plate is 98.3%, the precise location rate is 97.1%, and the average location time is less than 0.1 s. A better license plate location effect and accuracy are achieved by the proposed method.
-
表 1 车牌边缘颜色对
Table 1. Edge-color pairs of vehicle license plate
车牌 蓝牌 黄牌 白牌 黑牌 边缘颜色对 (1, 4) (2, 5) (3, 4)、(4, 5) (4, 5) 
下载: 导出CSV
-
[1] WANG Fei-yue, BROGGI A, WHITE C C. Road to transactions on intelligent transportation systems: a decade's success[J]. IEEE Transactions on Intelligent Transportation Systems, 2009, 10(4): 553-556. doi: 10.1109/TITS.2009.2033953 [2] BUCH N, VELASTIN S A, ORWELL J. A review of computer vision techniques for the analysis of urban traffic[J]. IEEE Transactions on Intelligent Transportation Systems, 2011, 12(3): 920-939. doi: 10.1109/TITS.2011.2119372 [3] 郑成勇. 一种RGB颜色空间中的车牌定位新方法[J]. 中国图象图形学报, 2010, 15(11): 1623-1628. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201011012.htmZHENG Cheng-yong. A novel license plate location method on RGB color space[J]. Journal of Image and Graphics, 2010, 15(11): 1623-1628. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB201011012.htm [4] 刘广起, 郑晓势, 张晓波. 基于图象纹理特征提取的车牌定位算法[J]. 中国图象图形学报, 2005, 10(11): 1419-1422. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB200511019.htmLIU Guang-qi, ZHENG Xiao-shi, ZHANG Xiao-bo. License plate location based on texture characteristic of image[J]. Journal of Image and Graphics, 2005, 10(11): 1419-1422. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-ZGTB200511019.htm [5] 陈振学, 常发亮, 刘成云. 基于特征显著性的多特征融合车牌定位算法[J]. 控制与决策, 2010, 25(12): 1909-1912, 1916. https://www.cnki.com.cn/Article/CJFDTOTAL-KZYC201012031.htmCHEN Zhen-xue, CHANG Fa-liang, LIU Cheng-yun. Multifeatures fusion license plates locating algorithm based on feature salience[J]. Control and Decision, 2010, 25(12): 19091912, 1916. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-KZYC201012031.htm [6] 朱成军, 蒲菊华, 高磊, 等. 基于Tophat变换和文字纹理的车牌定位算法[J]. 北京航空航天大学学报, 2008, 34(5): 541-545. https://www.cnki.com.cn/Article/CJFDTOTAL-BJHK200805015.htmZHU Cheng-jun, PU Ju-hua, GAO Lei, et al. License plate location based on Tophat transformation and text texture[J]. Journal of Beijing University of Aeronautics and Astronautics, 2008, 34(5): 541-545. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-BJHK200805015.htm [7] 刘纪红. 基于小波分析的车辆牌照定位方法[J]. 仪器仪表学报, 2002, 23(增3): 206-207. https://www.cnki.com.cn/Article/CJFDTOTAL-YQXB2002S1091.htmLIU Ji-hong. Wavelet analysis based approach for vehicle license plate locating[J]. Chinese Journal of Scientific Instrument, 2002, 23(S3): 206-207. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-YQXB2002S1091.htm [8] 任俊, 黄丹丹, 李志能. 结合纹理分析和支撑矢量机的汽车牌照定位研究[J]. 浙江大学学报: 工学版, 2006, 40(8): 1352-1357. doi: 10.3785/j.issn.1008-973X.2006.08.014REN Jun, HUANG Dan-dan, LI Zhi-neng. License plate locating using support vector machines and texture analysis[J]. Journal of Zhejiang University: Engineering Science Edition, 2006, 40(8): 1352-1357. (in Chinese). doi: 10.3785/j.issn.1008-973X.2006.08.014 [9] 章品正, 王健弘. 一种应用机器学习的车牌定位方法[J]. 应用科学学报, 2011, 29(2): 147-152. https://www.cnki.com.cn/Article/CJFDTOTAL-YYKX201102009.htmZHANG Pin-zheng, WANG Jian-hong. Vehicle license plate location based on machine learning[J]. Journal of Applied Sciences, 2011, 29(2): 147-152. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-YYKX201102009.htm [10] 盛曦, 吴炜, 杨晓敏, 等. 一种基于AdaBoost的车牌定位算法[J]. 四川大学学报: 自然科学版, 2007, 44(3): 535-539. https://www.cnki.com.cn/Article/CJFDTOTAL-SCDX200703017.htmSHENG Xi, WU Wei, YANG Xiao-min, et al. A license plate location method based on AdaBoost algorithm[J]. Journal of Sichuan University: Natural Science Edition, 2007, 44(3): 535-539. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-SCDX200703017.htm [11] FREUND Y, SCHAPIRE R E. A decision-theoretic generalization of on-line learning and an application to boosting[J]. Journal of Computer and System Sciences, 1997, 55(1): 119-139. doi: 10.1006/jcss.1997.1504 [12] FREUND Y, SCHAPIRE R E. A short introduction to boosting[J]. Journal of Japanese Society for Artificial Intelligence, 1999, 14(5): 771-780. [13] ZHANG C, CHEN T. An active learning framework for contentbased information retrieval[J]. IEEE Transactions on Multimedia, 2002, 4(2): 260-268. doi: 10.1109/TMM.2002.1017738 [14] TRAPEZNIKOV K, SALIGRAMA V, CASTANON D. Active boosted learning(ActBoost)[J]. Journal of Machine Learning Research, 2011(15): 743-751. [15] VIOLA P, JONES M J. Robust real-time face detection[J]. International Journal of Computer Vision, 2004, 57(2): 137-154. doi: 10.1023/B:VISI.0000013087.49260.fb [16] 李文举, 梁德群, 张旗, 等. 基于边缘颜色对车牌定位新方法[J]. 计算机学报, 2004, 27(2): 204-208. doi: 10.3321/j.issn:0254-4164.2004.02.008LI Wen-ju, LIANG De-qun, ZHANG Qi, et al. A novel approach for vehicle license plate location based on edge-color pair[J]. Chinese Journal of Computers, 2004, 27(2): 204-208. (in Chinese). doi: 10.3321/j.issn:0254-4164.2004.02.008 -
点击查看大图
图(4) / 表(2)
计量
- 文章访问数: 986
- HTML全文浏览量: 203
- PDF下载量: 2084
- 被引次数: 0
