Anti-Dust Ore Block Size Screening and Dynamic Warning Vision System.

Saved in:

Bibliographic Details
Title:	Anti-Dust Ore Block Size Screening and Dynamic Warning Vision System.
Authors:	Huang, Guangzhuang¹ 1912592749@qq.com, Xu, Shaochuan² shaochuanxu@163.com, Deng, Chao¹ 2358652439@qq.com, Yu, Xinzhong¹ 1953496819@qq.com, Ma, Minghao¹ mmh13591920408@163.com
Source:	Engineering Letters. Mar2026, Vol. 34 Issue 3, p867-880. 14p.
Subjects:	Image segmentation, Multiple target tracking, Dust control, Kalman filtering, Mine safety, Object recognition (Computer vision)
Abstract:	In the scene of underground mine crusher, highconcentration dust and ore stacking occlusion often cause edge blur and unstable tracking, limiting the reliability of traditional ore block detection and safety warning. To address these issues, this paper proposes an intelligent ore early warning system integrating improved instance segmentation and multitarget tracking. Based on the You Only Look Once version 11 nano framework, a dust-resistant context enhancement network is constructed. By introducing Switchable Atrous Convolution and Large Kernel Attention, global semantics and local features are collaboratively modeled to improve feature discrimination in complex environments. Meanwhile, doublelayer 3×3 convolution enhances edge detail expression, and the cross-scale feature fusion method is optimized to boost segmentation robustness and feature expression accuracy. At the tracking level, the system adopts the Kalman prediction and appearance feature matching mechanism of Deep Cosine Metric Learning with Simple Online and Realtime Tracking to achieve stable tracking of large ore and dynamic early warning for hazardous areas. Equipped with a visual interface, the system can display ore area parameters and trajectory status in real time. This study provides an intelligent scheme for underground ore block screening and safety early warning, which has application value for mine intelligence. [ABSTRACT FROM AUTHOR]
	Copyright of Engineering Letters 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:	In the scene of underground mine crusher, highconcentration dust and ore stacking occlusion often cause edge blur and unstable tracking, limiting the reliability of traditional ore block detection and safety warning. To address these issues, this paper proposes an intelligent ore early warning system integrating improved instance segmentation and multitarget tracking. Based on the You Only Look Once version 11 nano framework, a dust-resistant context enhancement network is constructed. By introducing Switchable Atrous Convolution and Large Kernel Attention, global semantics and local features are collaboratively modeled to improve feature discrimination in complex environments. Meanwhile, doublelayer 3×3 convolution enhances edge detail expression, and the cross-scale feature fusion method is optimized to boost segmentation robustness and feature expression accuracy. At the tracking level, the system adopts the Kalman prediction and appearance feature matching mechanism of Deep Cosine Metric Learning with Simple Online and Realtime Tracking to achieve stable tracking of large ore and dynamic early warning for hazardous areas. Equipped with a visual interface, the system can display ore area parameters and trajectory status in real time. This study provides an intelligent scheme for underground ore block screening and safety early warning, which has application value for mine intelligence. [ABSTRACT FROM AUTHOR]
ISSN:	1816093X