Research on Reading Recognition Technology of Gas Meter Based on Key Feature Matching

Abstract

In this article, we proposed a reading recognition technology for gas meters that can greatly substitute for artificial reading. The proposed technology includes three phases of work, which are image preprocessing, character location and segmentation, and character recognition respectively. The location method of color background and the vertical projection feature are employed to locate and segment characters, while the closed curve detection method, the template matching method and the pass line detection method are utilized and incorporated in character recognition phase. Finally, an experiment is conducted containing reading images of 100 samples with different angles and illumination, and experimental results verify the effectiveness of the proposed method.