A Meter Pointer Detection Method Based on Window Function Scanning

Abstract

The pointer meter is widely used in industrial production, and its data mainly relies on manual acquisition with low efficiency. The automatic recognition algorithm of pointer meter reading based on meter image is a hot research topic. In this paper, aiming at the complex environment of open-air pointer meter and the low stability of existing meter reading recognition algorithm, a method of pointer detection based on window function scanning is proposed. Firstly, the meter image is taken from the scene by template matching method. Separate and enhance the meter image based on the improved homomorphic filtering method; then construct a window function and scan the binary image of meter by the circumferential scanning method to find the position of the pointer. Experiments show that the pointer detection method proposed in this paper has better adaptability and stability under complex illumination environment.