结合HSV与局部Hough变换的指针式仪表识别算法

Pointer recognition algorithm based on HSV and partial Hough transformations

随着电力系统行业对自动化程度的要求不断提高,关于指针式仪表识别技术有待进一步改进。传统检测常常忽视阴影的影响,速度及准确性也常常不能满足要求,文章提出一种实用性强的识别算法。首先通过HSV空间变换操作消除阴影对指针识别的影响,结合减影法得到指针图像,然后利用Canny算法实现指针边缘的检测,再通过消除指针两端复杂图像以实现局部区域的Hough变换,最终利用指针夹角判别指针读数。实验结果表明,本算法的阴影消除效果明显,识别结果的准确性、实时性得到了较大提高。

For the increasing demands of automation in power system industry, the identification technology about pointer instrument should be improved further. Interference effects such as the shadow are often ignored in traditional detection methods. Here a more practical, accurate and quick recognition algorithm is proposed. First of all, through transforming all data into HSV space, influence of the shadow could be eliminated; and then, the pointer could be obtained by using subtraction method. Secondly, an edge image of the pointer could be obtained through canny edge detection method. Finally, by eliminating the impact of complex images at both ends of the pointer, and then applying the Hough transform, the reading of the pointer could be identified correctly. Experimental results show that it is efficient to eliminate the shadow impact with this algorithm, and the identification accuracy and speed have been improved greatly.