基于彩色区域识别快速报警指针识别算法.docxVIP

基金项目：华北电力大学大学生创新实践项目 基于彩色区域识别快速报警指针识别算法 刘薇，苏俊源，尚秋峰 （华北电力大学，河北 保定 071003） 摘要 针对电力监测系统中使用的彩色仪表，提出了一种基于彩色区域识别的快速报警指针识别算法。传统直线检测采用hough变换，具有存储空间大，计算时间长等缺点。创新性地将图像彩色信息和hough变换相结合，通过识别指针所在的区域颜色，判断是否需要报警输出，并为下一步精确识别指针示数确定hough变换的搜索范围，通过减小搜索范围来减少hough算法需要的内存空间和计算时间。实验结果表明，，改进算法比传统算法具有明显更高的效率。 关键词 ： 彩色区域识别；Hough变换 ；快速报警；指针识别 中图分类号：TM933 文献标识码：A 文章编号：1001-1390（2014）20-0000-00 A Fast Alarm Pointer Recognition Algorithm Based based on Color Region Recognition LIU Wei, SU Jun-yuan, SHANG Qiu-feng （North China Electric Power University, Baoding 071003，HeiBei，China） Abstract：A fast alarm pointer recognition algorithm based on color region recognition has been proposed in this paper Ffor the color instrument used in the electric monitoring system,this paper proposes a fast alarm pointer recognition algorithm based on color region recognition.The traditional pointer recognitionline detection algorithms take use of Hough transformation, which ask forwith the disadvantages of a large storage space and a long computation time for compution. This paper cCombinesing the image color information and Hough transformation creatively, whether the alarm signal should be outputted will be determined byby identifying the color of the region where the pointer lies, which will confirm the , judging whether to alarm output, and determining the search range of of the Hough transformation algorithm for the accurate next identification. The required memory space and computation time is will be reduced by reducing shrinking the search range. Experiment results show a higher efficiency of Tthe improved algorithm has efficiency significantly higher than that of the traditional algorithm. Key words: color region recognition, Hough transformation, fast alarm, pointer recognition. 0引 言 在电力系统中，指针式仪表因其结构简单、读数直观、操作容易、成本低等优点被广泛应用。这种仪表每天都需要记录大量的数据。目前国内记录仪表数据主要靠人眼观测，精度低、可靠性差、效率低、劳动强度大。因此，研发一种快速、准确、稳定可靠的指针式仪表的自动识别系统具有非常重要的意义。 在电力仪表自动判读系统中，由视频采集设备所采集到的现场图像经过保存后通常为彩色图像。彩色图像的信息量非常丰富，可以作为图