Research on Intelligent Instrument Character Location Technology Based on Computer Vision

The digital character recognition system of remote meter reading system is mainly composed of two key parts: character location and recognition. And accuracy of character recognition has much work with whether or not the characters to be recognized can be accurately located. The border stain and noise of character area cause the character to be recognized to locate difficultly. To address this reality, this paper presents an effective algorithm for character repositioning. Firstly, the image is preprocessed including gamma correction, defogging and denoising before location. Then the image is converted to binary data, the object contour is extracted, and the region of interest of the candidate character is determined. The region of interest of the character is determined again according to the angle of the candidate character's region of interest. Finally, the region of interest of the character is segmented and the character is repositioned to determine the final character to be recognized. The experimental results show that the proposed method is simple and effective, and the localization accuracy is very high, strong robustness. In addition, satisfactory results can be obtained in the actual environment.