Reading recognition method of analog measuring instruments based on improved hough transform

To effectively resolve the automatic recognition problem of analog measuring instruments, a reading recognition method based on improved Hough transformation is proposed. Firstly, the paper presents the image preprocessing operation. In order to make the image filtering better, an improved adaptive median filter method is adopted. Secondly, we explains in detail the recognition principle for the pointer angle using the improved Hough transform. And the pointer's reading is got using the linear relationship between scale and angle of instrument dial revolutionary. In order to further improve the recognition accuracy, as to the situation that pointer completely coincides with the whole tick, late correction will be done. Finally, the automatic recognition software is written in VC++6.0 IDE, and complete the recognition experiment for the AC voltmeter. The experimental results show that the method is very valid for the pointer angle recognition. In addition, the advantages of this study are as following: simple algorithm, good real-time and high precision.