Hough变换的改进及其在指针式水表识别过程中的应用  

The Improved Hough Transform and Its Application in Recognition on Pointer Meters

作　　者：李静[1,2] 张宁[2] 张岩[1,2] 时慧 

机构地区：[1]北京化工大学,北京100029 [2]北京石油化工学院,北京102617 [3]河北省计量监督检测院,河北石家庄050000

出　　处：《北京石油化工学院学报》2015年第1期58-64,共7页Journal of Beijing Institute of Petrochemical Technology

摘　　要：在图像处理中,Hough变换广泛地应用于图像空间图形的检测,是从图像中识别几何图形的基本方法之一,但是其对于除直线之外的检测基本都会涉及3个甚至3个以上的参数,导致变换过程中所使用的存储空间大,计算时间长。针对传统Hough变换的算法缺点,提出一种改进的Hough变换算法,首先对原图像进行灰度化、滤波等图像预处理,然后结合图像的结构信息缩小变换像素点所处的范围,从而大大减少计算时间。进一步再将改进的Hough变换应用于图像特征提取,即圆和指针的提取。最后根据圆心连线与指针的角度来对指针的读数进行判别。In image processing, the Hough transform is widely used in the detection of image space graphics, which is one of the basic methods to identify geometry from the image. But it ba- sically involves three or even more than three parameters except the straight line detection, which leads to the excessive use of storage space in the process of transformation and long computing time. To overcome the weak points of the traditional Hough transform algorithm, this paper proposes an improved Hough transform algorithm. First of all, the image pre-processing is conducted, including graying and filtering of the original image. Then the narrow scope of trans- formation pixel points is reduced by combining with the structural information of image, which greatly reduces the computing time. The improved Hough transform is applied to image feature extraction, which contains the extraction of the circle and pointer. At last, the reading can be in- terpreted according to the angle between the attachment of centers of the circles and pointer.

关 键 词：图像处理 HOUGH变换 指针式水表 角度判读 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]