无人机检测在车牌字符自动识别系统中的应用  

Application of UAV Detection in Automatic License Plate Character Recognition System

作　　者：陈洪亮 李雪松 关巍 孙同展 侯超 CHEN Hong-liang;LI Xue-song;GUAN Wei;SUN Tong-zhan;HOU Chao(Stae Grid Hebei Electric Power Co.,Ltd.,Xiong'an New Area Electric Power Supply Company,Xiong'an 071800 China;Tianjin Richsoft Electrjc Power Information Technology Co.,Ltd.,Tianjin 300000 China;Hebei Electric Power Co.,Ltd.,Shijiazhuang 050000 China)

机构地区：[1]国网河北省电力有限公司雄安新区供电公司,河北雄安071800 [2]天津市普迅电力信息技术有限公司,天津300000 [3]国网河北省电力有限公司,河北石家庄050000

出　　处：《自动化技术与应用》2022年第10期81-84,共4页Techniques of Automation and Applications

基　　金：国网河北省电力有限公司科技项目资助(B304XQ200010)。

摘　　要：传统车牌字符自动识别系统忽略了字符本体信息特征的提取,导致车牌字符识别误差较大。为此,提出基于无人机检测技术的车牌字符自动识别系统。构建车牌字符自动识别的无人机检测图像采集模型以及图像融合模型。通过边缘参数分割和融合特征识别法,实现无人机图像像素融合和特征分解。提取无人机采集图像资源的二元语义本体信息特征,结合图像的语义特征融合方法,实现对车牌字符的自动识别。并采用深度学习方法寻优车牌字符识别结果。仿真实验结果表明,所提的车牌字符识别方法可靠性较好,识别精度较高。The traditional license plate character recognition system ignores the extraction of character ontology information features, resulting in small license plate character recognition error. Therefore, the design method of license plate character automatic recognition system based on UAV detection technology is proposed. The UAV detection image acquisition model and image fusion model of license plate character automatic recognition are constructed. The pixel fusion and feature decomposition of UAV image are realized by edge parameter segmentation and fusion feature recognition. The binary semantic ontology information features of UAV image resources are extracted, and combined with the semantic feature fusion method of image, the automatic recognition of license plate characters is realized. The deep learning method is used to optimize the license plate character recognition results. Simulation results show that the proposed license plate character recognition method has good reliability and high recognition accuracy.

关 键 词：无人机检测 车牌字符 自动识别 图像滤波 本体信息特征 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]