基于空间变换的指针式仪表读数识别算法研究

针对目前指针式仪表图像自动读数识别普遍存在受光照变化影响大,表盘内部阻尼液干扰引起识别精度低的问题,本研究提出了一种基于空间变换的指针式仪表读数识别算法。首先选取指针所在区域的圆心和半径,获得边缘提取的仪表内圈圆弧轮廓,以此确定图像空间变换需要的参数从而将圆弧形的表盘转换成矩形;其次通过图像拼接处理构建无指针图像,在此基础上对刻度线进行校正处理得到刻度线坐标;最后通过减影法提取指针区域中心直线坐标,判断指针与刻度线的相对位置来获取指针读数。光照实验表明,本算法对指针和刻度线识别受光照影响小,经校表实验表明,本研究提出的算法识别速度快、精度高。

A Algorithm for the Recognition of Pointer Type meter Based on Space Transformation

At present,the auto reading recognition algorithms for pointer type meter are vulnerable to illumination,and the recognition accuracy for the meter with damping liquid is not high.In this paper,a new algorithm for the recognition of pointer meter based on space transformation is proposed.The proposed approach firstly selects the circle of meter region,and extracts the inner circle contour of the region by edge extraction in order to confirm the parameters of image transform and dials into the circular rectangle;Secondly,the pointer image is obtained by image mosaic processing,on this basis,it redraws the scale lines to get its coordinates;Finally,meter reading is derived by image subtraction to obtain the center line coordinates of the pointer region,and then,the reading recognition is identified by the relation between the pointer and scale lines.Light experimental results show that the identification of the pointer and scale lines are smaller effected by illumination.Calibration instrument results show that the proposed algorithm has better recognition efficiency and accuracy.