Adaptive Scale Recognition System for Pointer-Type Pressure Gauges in Industrial Scenes Based on Enhanced YOLOv11n-Pose and EasyOCR.

Saved in:

Bibliographic Details
Title:	Adaptive Scale Recognition System for Pointer-Type Pressure Gauges in Industrial Scenes Based on Enhanced YOLOv11n-Pose and EasyOCR.
Authors:	Deng, Chao¹ shaochuanxu@163.com, Xu, Shaochuan² 1912592749@qq.com, Huang, GuangZhuang³ 1953496819@qq.com, Yu, Xin Zhong¹ mmh13591920408@163.com, Ma, Ming Hao¹
Source:	IAENG International Journal of Computer Science. Mar2026, Vol. 53 Issue 3, p1248-1260. 13p.
Subjects:	Pressure gages, Optical character recognition, Industrial safety, Feature extraction, Wavelet transforms
Abstract:	The pointer-type pressure gauge serves as a crucial safety monitoring device during the operation of liquefied natural gas tank trucks, and its recognition accuracy is directly linked to the safety and accident prevention efficiency of the storage and transportation system. To address the limitations of traditional manual inspection methods--such as low detection efficiency and insufficient accuracy under complex operating conditions--this study proposes an automatic pressure gauge recognition system based on the integration of You Only Look Once version 11 Nano for pose estimation and Easy Optical Character Recognition. The system detects key points on the gauge dial, including the pointer center, pointer tip, zero scale, and end scale, and combines them with Easy Optical Character Recognition to achieve end-to-end automatic reading across multiple gauge types. To enhance detection capability under challenging environmental conditions, the proposed network incorporates Wavelet Convolution (Wavelet Transform Convolution 2D), Cross Stage Partial 3-layer block with Wavelet Convolution module, Feature Pyramid Network P2 small-target detection head, and lightweight Efficient Multi-scale Channel Attention attention mechanism into the network structure. Wavelet Convolution effectively fuses spatial and frequency-domain features to strengthen multi-scale feature representation, while Efficient Multi-scale Channel Attention performs cross-channel feature selection and noise suppression, significantly improving recognition performance under low illumination and long-distance imaging scenarios. Experimental results demonstrate that the system maintains high efficiency and precision even in complex environments. Field deployment further validates that the system achieves substantially higher detection efficiency than manual inspection, effectively mitigating tank truck congestion during peak operations and preventing potential safety incidents. [ABSTRACT FROM AUTHOR]
	Copyright of IAENG International Journal of Computer Science is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	The pointer-type pressure gauge serves as a crucial safety monitoring device during the operation of liquefied natural gas tank trucks, and its recognition accuracy is directly linked to the safety and accident prevention efficiency of the storage and transportation system. To address the limitations of traditional manual inspection methods--such as low detection efficiency and insufficient accuracy under complex operating conditions--this study proposes an automatic pressure gauge recognition system based on the integration of You Only Look Once version 11 Nano for pose estimation and Easy Optical Character Recognition. The system detects key points on the gauge dial, including the pointer center, pointer tip, zero scale, and end scale, and combines them with Easy Optical Character Recognition to achieve end-to-end automatic reading across multiple gauge types. To enhance detection capability under challenging environmental conditions, the proposed network incorporates Wavelet Convolution (Wavelet Transform Convolution 2D), Cross Stage Partial 3-layer block with Wavelet Convolution module, Feature Pyramid Network P2 small-target detection head, and lightweight Efficient Multi-scale Channel Attention attention mechanism into the network structure. Wavelet Convolution effectively fuses spatial and frequency-domain features to strengthen multi-scale feature representation, while Efficient Multi-scale Channel Attention performs cross-channel feature selection and noise suppression, significantly improving recognition performance under low illumination and long-distance imaging scenarios. Experimental results demonstrate that the system maintains high efficiency and precision even in complex environments. Field deployment further validates that the system achieves substantially higher detection efficiency than manual inspection, effectively mitigating tank truck congestion during peak operations and preventing potential safety incidents. [ABSTRACT FROM AUTHOR]
ISSN:	1819656X