复杂仪表的图像自动读数识别方法

A Method of Automatic Indication Recognition for a Multi-Pointer Meter by Image Processing and Recognizing

在对复杂仪表的特点分析研究的基础上,提出了复杂仪表的图像自动读数识别的原理和方法。对实现其原理所采用的主要方法,如分段线性灰度变换、Hough变换、模板匹配法进行了介绍。以水表为实例证明了所采用的图象自动读数识别原理和方法的正确性和有效性。结果表明,该方法具有很好的稳定性、准确性、快速性等特点。

A method of automatic indication recognizing for a multi-pointer meter such as a water meter by image processing and recognizing is presented. The principle and step of the automatic indication recognition meter are described. Image analysis and recognition algorithms, such as piece wise-linear gray level transformation and Hough transformation, and the template matching method, are used synthetically to automatically take out the direction and position of the pointers on the meter panel and to determine the meter's indication from the image of the meter panel. It is proven to be effective to recognize the water meter's indication by this method. The result indicates that this method is reasonable and applicable.