Study On Smart Electric Meter Error Detection Methods Based On Image Recognition Technology

Watt-hour meter is the only device that is used to measure the amount of consumed electricity, preventing the occurrence ofelectricity stealing is beneficial to reducing the grid economy loss, raising the level of power grid safe operation, saving costs and reducing labor intensity and protecting the interests of users, etc.Given that the number of users on the low voltage side is very large, power stealing techniques are diverse, professional testing equipment is expensive and the traditional inspection methods is rather complicated, studying a fast method for powerstealing check is very important.Current research mainly focuses on image processing and the measuring error factors of watt-hour meter. This articlestudies a fast power examination method which takes watt-hour meter pulse modulationas the research object. By means of pulse modulationimage processing and watt-hour meter measuring error detection software, the watt-hour meter measuring power value can be achieved, and then the results is compared with the actual measured valuesto judge whether there is power stealing.Firstly, this paper studies the theory of watt-hour meter pulse light image recognition technology. Image enhancement, color image processing, image morphology filtering, image target detection, etc, are introducedrespectively. Then combined with the related theories and the actual image processing,watt-hour meter pulse modulation recognition process is given, and the characteristics of the brightness feature, color feature and shape featureare studied in both the general situation and the special circumstances.Secondly, based on pulse modulation image recognition technology,Android watt-hour meter measuring error detection software is programmed. The origin of the Android, application components, the setting up of development environment, the structure of the source files and so on are introduced in detail.Then combinedwith OpenCV software API, the software is designed and developed, and the related contents of software is discussed detailedly.Finally, the precision analysis and application promotion is conducted. At first, the testing precision of the measurement software is tested, and the detection method is compared to the traditional test method to illustrate the advantages of our method.Then the application examples are given, the application prospect of test method is expounded.