Pointer-type meter automatic reading from complex environment based on visual saliency

Authors:

Abstract
This paper presents an automatic interpretation method for pointer-type meter reading from complex environment based on visual saliency. Compared with the manually based methods, the proposed approach is based on image processing and automatic technique, avoiding the problems of low efficiency, high subjective, as well as inaccurate precision that are caused by manual judgement. Firstly, a localized pixel inhomogeneity factor (LPIF) model is developed to improve the visual saliency of the pointer-type meter region. After binarization, the Hough circle detector is utilized to locate the approximate pointer region. In the candidate image region, the true pointer-type metre indicator is considered as the straight line with maximum length. Therefore, the Hough line transform is used to locate the precise pointer-type metre indicator. In the final step, the scale of metre-type meter reading is obtained by calculating the angle of pointer indicator. Experiments demonstrate that the proposed method is very effective on all kinds of pointer-type meter readings. And the proposed automatic scheme not only greatly economizes the labor cost, but also deals with life-threatening issues in dangerous environment.