Calculus accumulation-based method for extracting values of instrumentation image

• Published in: Journal of Applied Science and Engineering Vol. 29; no. 2; pp. 257 - 265
• Main Authors: Dianming Wang, Li Zhou, Xiangxin Chen, Mingxin Yi, Xiaoju Yin
• Format: Journal Article
• Language: English
• Published: Tamkang University Press 2026
• Subjects: calculus accumulation; dual needle gauge recognition; two dimensions convolution; two-axis coordinate system
• ISSN: 2708-9967; 2708-9975
• DOI: 10.6180/jase.202602_29(2).0002

The working conditions of outdoor pointer instruments are complex, and the dial is vulnerable to environmental pollution, so it is difficult to identify the readings online accurately. A double-pointer meter reading recognition scheme based on two-dimensional convolution and calculus accumulation is proposed. The weighted multifeature matching algorithm is used to accurately obtain the center coordinates of the rotating shaft to improve the success rate of meter image recognition. The template filtering method eliminates cross-pointer interference. The polar coordinate transformation and calculus accumulation of the dial image are carried out to obtain the position of the target double-pointer and realize automatic reading, which is suitable for most instruments. Through the experimental comparison and analysis of outdoor dust occlusion, droplet occlusion, leaf occlusion, and different lighting conditions, the results show that the algorithm has good robustness, is almost not disturbed by the external environment, and has high recognition accuracy. The recognition success rate is more than 98 %, which provides a good reference for the unattended operation of pointer instruments.