Abstract: In order to realize the automatic reading of the pointer meters, a reading method based on improved ORB and Hough algorithm for pointer meters is proposed in this paper. Firstly, the corner point enhancement method is used to enhance the keypoints detected by the ORB algorithm, and the perspective transformation matrix between the template image and the image to be detected is calculated by the feature matching points. Then, the pointer is extracted by image preprocessing, such as mathematical morphology processing and threshold segmentation. A similar feature triangle method based on feature matching point pair is put forward to determine the rotation center of the pointer. Combining the projection method to judge the pointer direction, the pointer refinement algorithm and the Hough transform algorithm with the center constraint are used to detect the pointer angle. Finally, the reading results can be obtained according to the priori meter information. The experimental results show that the algorithm can meet the requirement of pointer meter reading in terms of reading speed and precision, and it has high reliability and engineering application value.

Key words: pointer meter, Oriented FAST and Rotated BRIEF（ORB）, Hough transform, pointer extraction, center extraction

摘要： 为了实现指针式仪表的自动识读，提出一种基于改进ORB（Oriented FAST and Rotated BRIEF）和Hough变换算法的指针式仪表识读算法。利用角点强化方法加强ORB算法检测的特征点，通过特征点匹配对计算模板图像与待检测图像之间的透视变换矩阵。利用数学形态学处理、阈值分割等图像预处理提取指针，并提出一种用于确定指针旋转圆心的基于ORB特征匹配对的相似特征三角形方法，结合投影法定位指针方向。利用指针细化算法和添加圆心约束的Hough变换算法检测指针角度。最后根据仪表的先验信息得到读数结果。实验结果表明该算法在识读速度和精度等方面都能够满足指针式仪表识读的要求，具有较高的可靠性和工程应用价值。

关键词: 指针式仪表, ORB, Hough变换, 指针提取, 圆心提取