Research on Pre-processing for Pointer Meter Automatic Recognition

摘要

The accurate detection of the pointer position is the key of the automatic pointer meter reading. Appropriate image preprocessing is the basis of the precise location of the scale zone and pointer on the dial. In this paper, the images are preprocessed by image grizzle, median filter and image binarization. In order to improve the quality of the image segmentation, an improved bigness method based on gray stretch is adopted in the threshold segmentation. Experiments results show that effective thresholds are obtained, the useful information is successfully reserved and satisfactory segmentation results of the pointer meter images are achieved.