Study of low-contrast digital meter image segmentation algorithm

Abstract

During the digital recognition process of digital display instrument, if there is no lighting or light is weak, the recognition rate comes low because of poor image quality. As to this problem, this paper proposes a new algorithm for image segmentation. The algorithm uses the histogram equalization to improve image contrast and enhance gray difference between the background and the area to be identified. And then it positions display area coarsely using the adaptive region growing method. Finally, the display area is precisely positioned by projection method. And adaptive threshold segmentation is done in this area. The experimental results show that the algorithm can rapidly and accurately position the display area of digital display instrument. And segmentation effect of the positioned image is improved obviously, effectively simplifying the subsequent digital recognition algorithms and guaranteeing its recognition accuracy.