【作者】 戴亚文；

【导师】 王三武；

【作者基本信息】 武汉理工大学 ， 机械制造及其自动化， 2003， 硕士

【摘要】 对于大多数自来水公司，都是依靠人工抄表进行用水和收费管理。而且在抄表过程中由于抄表不到位、私自调表改读数等人为因素造成的水损是全部水损的主要部分。为便于管理，降低水损，市场上出现一种带单片机的电子水表。但存在水表更换和管线改造的问题，使用成本高，且电子水表易进水造成水表失效，使用可靠性差。如何避免如此种种弊端成为一大难题。随着计算机图像处理技术的不断发展，我们提出了应用数字图像处理技术开发水表读数自动识别系统实现水表自动抄表和自动识别读数的解决方案，并成功开发出数字图像水表读数自动识别系统。 将图像处理技术应用于水表多指针的读数识别，是一种新方法，国内目前尚无相应的开发。本文结合水表的具体特点，分析了将图像处理技术应用于自动识别水表指示读数的原理。找到了一种满足一定速度和精度的水表读数识别方法。该方法主要包括水表图像预处理、水表图像分割、指针读数的识别。在预处理中提出了一种基于灰度期望的分段线性变换；对水表的图像分割，提出基于面积和结构信息的二值化自动分割，包括基于结构和灰度特征的哈夫变换分割法(指针信息区的提取)和基于指针面积的特征的阈值分割(指针提取)；针对指针图像的非均匀光照、背影等干扰，提出基于多特征的分区模板匹配法对指针的方向进行识别，从而计算出读数。 利用本文的算法开发的数字图像水表读数自动识别系统对现场采集的大量水表图像进行测试，结果表明该方法抗干扰能力强，识别率完全达到要求。该系统为自来水公司实行计算机管理提供了新的技术手段。更多还原

【Abstract】 For the majority of tap water companies, the record and charge management of the water consumption is all to depend on the manual methods. To improve the present backward condition, an automatic reading system of water meter with CCD has been developed so that water meter can be automatically read by applying image processing technique, which is completely new in our country at present.This paper, after discussing the basic theory and the application of digital image processing and analyzing the concrete characteristics of the water meter, puts forward the crux algorithm that identifies automatically the position and direction of the pointers of water meter with techniques of the image preprocessing, image segmentation and image analysis and recognition. In the image preprocessing that includes image smoothing and linear gray level transformation, the author brings forward a mended linear gray level transformation based on mean of gray value to enhance the image contrast. In the segmentation the paper presents an arithmetic taking the area of the pointers and construction feature of the water meter into consideration, which can extract the part of pointers in the special space of the water meter, and then binarizes image using a dynamic threshold based on the area of the pointer. The design can enhance the recognition rate for the next step. Finally, considering the optics of the nonuniform shining on the water meter and the shadow jam of pointers, the author employs the mold plate-matching method based on the several characteristic parameters of the pointers of the meter to recognize the indicating values of the pointer, The characteristic parameters include the area, the gray, the shape and the information of distribution of the pointer’s gray.These series of processing methods offer high stability and goodrepetition. Abundant experiments have proved that this system gives a feasible method in the recognition of the reading of the water meter. The recognition rate of this system has reached the request.更多还原