基于图像处理的细指针表盘识别方法研究

Research on the Recognition Method of the Fine Pointer Dial based on Image Processing

摘要:

为了解决在细指针表盘识别中难适应光照变化、细指针难分割和表盘读数准确识别率低的问题，结合细指针表盘区域的图像特点，采用图像处理对细指针表盘读数进行了研究。首先，对表盘进行标定，获得表盘量程等先验数据，制作模板，并提出了基于SURF算法改进的多任务模板特征匹配算法；其次，根据指针活动区域，提出了基于分割扇形区域的指针定位法(Pointer positioning method based on Segmented Sector, PSS)，进行图像预处理和二值化，使用最小二乘法拟合求得角度；最后，由指针角度得到读数。实验结果表明，该方法对细指针表盘识别率高，能够准确、稳定地识别细指针表盘示数。

Abstract:

In order to solve such problems that difficult adaptation of the illumination change during the recognition on the fine pointer dial, difficult segment of the fine pointer and low accuracy of the dial reading recognition , combined with the image characteristics of the fine pointer dial area, a image processing method is proposed to calculate the reading of the fine pointer dial. Firstly, the dial is calibrated, the prior data such as the dial range is obtained, the template is created, and the multi-task template feature matching algorithm based on SURF algorithm is proposed. Secondly, according to the pointer active area, a pointer positioning method based on the segmented sector (PSS) is proposed to locate the pointer. Through the image preprocessing and binarization, the angle is calculated using the least squares method. Finally, the reading is obtained based on the pointer angle. The experimental results show that the presented method has a high, accurate, and stable recognition rate for the fine pointer dial.