基于改进霍夫-极坐标变换的圆形指示表表盘刻度自适应识别算法OACHSSCD

Aims:Aiming at the difficulties in the reading recognition of circular indicator gauge dial images caused by uneven scale distribution,tiny pointer structure,blurred edges,low contrast and uneven illumination,this paper proposed a high-precision automatic reading method.Methods:Firstly,Gaussian filtering was employed to preprocess the images so as to suppress noise and enhance edge features.Then,the Hough circle transform was used to preliminarily locate the dial region;and the least squares method was adopted to optimize the parameters of the circle center and radius.The circular dial was converted into a rectangular image through polar coordinate transformation,which simplified the geometric relationship between the pointer and the scales.Finally,Otsu adaptive threshold segmentation,image inversion and the Canny operator were used to achieve feature enhancement and edge detection;and the pointer angle was extracted via column-direction projection to realize automatic reading.Results:Experiments of 6 instruments gauge images showed that the average absolute error was 0.10 and the average relative error was 1.46％.Conclusions:The results demonstrate that the proposed method has high recognition accuracy and good engineering application value.