Lightning Arrester Monitor Pointer Meter and Digits Reading Recognition Based on Image Processing

Lightning arresters, also known as surge arresters, are used to protect electrical equipment from high-transient overvoltage damage. Lightning arrester monitor is a meter that combines pointer with digits, which can be used to detect leakage current and action (discharge) times. This paper proposes a simple meter image processing approach for obtaining the images at a fixed position and record the reading. First, the image is preprocessed, including noise reduction, grayscale processing and histogram equalization. The canny operator is then used to get a binary image containing the edge information. Finding circle and center to match the scale area by using circumscribed Hough transform, where can be obtained by eliminating unnecessary information by morphological processing. The standard Hough transform is used to confirm the pointer, and the result of the pointer meter reading can be recognized. About the digits reading, the horizontal and vertical projection (HVP) method is used for twice to determine the counter area and segment characters. The counter reading is acquired by template matching.