Water meter digital timing recognition system based on edge calculation

Abstract:
Water is one of the indispensable resources in human production and life, so the accurate management and measurement of water resources is also a very important work. According to statistics, many regions in China still use manual meter reading for water meter measurement and water management. Although this method is simple and easy, it is a huge task for staff. In addition, due to the influence of people's physiological state and emotional factors in the process of meter reading, it is easy to cause mistakes and omissions, which has been difficult to meet the current actual needs. In view of the problems of missing and wrong acquisition caused by manual meter reading in many water plants and communities at present, a water meter timing recognition method based on edge computing is proposed, and an unmanned meter reading system using edge computing and computer vision technology is established to realize information-based water resource management, which meets the needs of social development. Firstly, the digital region of the water meter is located by Canny edge detection and Hove line detection, and then the located digital image is Gaussian filtered; Then build the Jetson Nano compilation environment, and finally deploy the YOLO-v3 detection network to the Jetson Nano computing board to complete the digital timing recognition in the water meter. The experimental results show that the detection time of the water meter in the Jetson nano is 8-9s, which meets the actual needs. This method is feasible and effective.