基于数字图像处理技术的双指针机械仪表读数自动识别方法

A Method for Automatic Recognition of Dual-pointer Instrument Reading Based on Digital Image Process Technology

针对工程应用中大批量双指针仪表的读数识别问题,提出了一种基于数字图像处理技术的自动识别算法:首先对图像进行预处理,包括图像分割、滤波及膨胀腐蚀等步骤;然后设计了一种限定阈值的Hough变换,确定不同指针的位置关系,并根据指针与旋转中心的距离关系标定不同指针的矢量信息;最后利用表盘刻度与指针夹角的线性关系计算指针读数,并对指针与正刻度线重合、不同指针反向等情况进行校正处理。试验结果表明,该算法对于均匀刻度的双指针仪表有出色的测试精度和识别效率。

Aiming at the problem of reading recognition for quantities of dual-pointer instruments in engineering application,an automatic recognition algorithm based on digital image processing is proposed.Firstly,the image is preprocessed such as image segmentation,filtering,dilation and erosion et al.Then,a threshold limited Hough transform is designed to detect the location of different pointers.And the vector information of different pointers is labeled according to distance relationship between pointer and the center of rotation.Finally pointers' reading is obtained by using the linear relationship between scale and the pointer angle,the pointer is coincide with the positive calibration line and corrected.The test result shows the algorithm is very practical for reading recognition of uniform scale pointers,and has excellent test accuracy and recognition efficiency.