摘要
交通信号数字指示灯识别是根据数字指示灯的特征分析图像内容,找出数字指示灯目标并加以分类识别的过程,包括预处理、定位、分割以及识别等过程。本文采用基于颜色-形状特征的目标检测算法和基于BP神经网络的识别算法,设计一种较为简单、准确的数字信号灯的识别方法。仿真实验表明,本算法具有较高的准确率和执行效率。
Identification of digital traffic signals indicator is the process of analyzing image content based on the characteristics of digital indicator to identify, classify and recognize digital indicator target, including pretreatment, positioning, segmentation, i- dentification and etc. This paper designs a more simple, accurate digital signal light recognition method by using the target detec- tion algorithm based on the color-shape characteristics and the recognition algorithm based on BP neural network. Simulation ex- periments results show that this method is of higher accuracy rate and execution efficiency.
出处
《计算机与现代化》
2013年第4期77-80,共4页
Computer and Modernization
基金
江苏省自然科学基金资助项目(BK2011375)
关键词
BP神经网络
信号灯识别
数字指示灯
BP neural network
semaphores recognition
digital indicator