一种新的基于图像识别技术的信号灯识别算法  

A New Algorithm for Signal Lamp Recognition Based on Image Recognition Technology

作　　者：王鹏[1] 郑光宇[1] 宋开亮[1] 

机构地区：[1]第二炮兵工程学院302教研室,陕西西安710025

出　　处：《兵工自动化》2009年第3期73-75,共3页Ordnance Industry Automation

摘　　要：将远程数字视频监控和图像识别技术引入武器系统的测试过程中。信号灯作为武器系统测试中最常见的识别对象,多在灰度空间和RGB颜色空间中进行识别,但该两种方法均存在一定不足。鉴于HSI颜色模型在图像信息描述中的优越性,引入HSI空间颜色对比相似距离,用以改善信号灯的识别效果。新算法在RGB空间中进行均值的求解,再转化为HSI颜色,可以减小计算量,提高效率。实验表明,新算法实用性强,有效性高。The paper proposes a new algorithm, which introduces remote digital video monitoring and image recognition technology into the testing process of weapon system. As the most common recognition object in weapon system testing, the recognition of signal lamp always be implemented in gray space or RGB color space, but both of the methods have shortages. Considering HSI color space＇s advantage in describing image information, HSI color pair＇s similarity measure was introduced to improve the recognition effect. In the new algorithm, the average color was solved in RGB color space, and then conversed to HIS color; the quantity of calculation was reduced and efficiency was advanced. Experiments show that the new algorithm has strong practicality and high validity.

关 键 词：图像识别 信号灯 颜色空间 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP301.6[自动化与计算机技术—计算机科学与技术]