摘要
在当今人工智能时代下,车标属于车辆主要标识,但需要能准确的识别,难度就较高,如何利用车标识别技术,来解决当今的智能交通中的诸多问题,是值得思考的。基于深度学习的CNN模型的车标识别方法能提高对车标识别的准确度,将车标当做探测对象,从而提高车标识别的精确性,这样解决了有关车辆辨别问题,深度学习下的车标识别方法可以通过多样化的特征学习,能直接从训练样本集中提炼出特征。
In today's era of artificial intelligence,as an important vehicle logo logo, can accurately identify the high degree of difficulty, can solve the traffic problems existing in intelligent transportation system using logo recognition technology. Vehicle logo recognition method under CNN deep learning model can improve the accuracy of vehicle recognition, the logo as a research object, to improve the accuracy of recognition, so as to solve a series of problems. Vehicle logo recognition method under deep learning through diversified learning, can directly extract the features from the training samples, the logo figure get into the neural network classifier for classification.
出处
《电脑知识与技术》
2018年第4Z期207-209,共3页
Computer Knowledge and Technology
基金
贵州省2017年大学创新创业训练计划项目"基于深度学习CNN模型的智能识车助手"(项目编号:201714223042)
"贵州师范学院大学生互联网+创新创业训练中心"(项目编号:黔教高发[2015]337号
黔教高发[2017]158号)
贵州省高技术产业示范工程专项项目(黔发改投资[2015]1588号)
贵州省教育厅创新群体重大研究项目(合同编号:黔教合KY字[2016]040)
贵州省普通高等学校工程研究中心(合同编号:黔教合KY字[2016]015)
关键词
车标识别
深度学习
CNN模型
智能交通
人工智能
recognition
deep learning
CNN model
intelligent transportation
artificial intelligence