Article Details

Abstract

To address the problems of low manual inspection efficiency, delayed hidden danger identification and difficult monitoring of high-risk areas in the safety supervision of industrial and livelihood scenarios such as quarries, open-pit mines and reservoirs, this study proposes a multi-target recognition system based on improved YOLOv11 to realize the intelligent identification and management of all elements of "human, machine, environment and management". The CBAM attention mechanism is introduced to optimize the feature extraction network, K-means++ adaptive anchor box clustering is adopted to improve the detection accuracy of small targets, and the CIoU loss function is combined with a mixed data augmentation strategy to optimize the model training process, thereby constructing a target recognition framework suitable for complex industrial environments. Verification results in two typical scenarios, open-pit mines and reservoirs, show that the system achieves an average recognition accuracy of 92.3%, a recognition rate of no less than 85% for cracks of 20 mm and above, and a response time of no more than 10 s for abnormal targets. The system can meet the demand for 24/7 uninterrupted safety supervision and provides technical support for the digital transformation of safe production in industrial scenarios.

Keywords

YOLOv11; multi-target recognition; safety supervision; industrial scenario; feature enhancement; human-machine-environment-management

References

1. [1] Li Yun, Xu Wei, Wu Wei Application and Research of Unmanned Aerial Vehicle Technology for Disaster Monitoring [J]. Disaster Science, 2011, 26 (1): 138-143
2. [2] Hu Jiansong, Wang Feijiang, Qiu Benfang, etc Research on Key Technologies and Applications of Unmanned Aerial Vehicle Engineering Inspection Based on Deep Learning [J]. Shanxi Construction, 2026, 52 (1): 126-132
3. [3] Wang Zhebing, Cao Hongxia, Zheng Wen Low altitude autonomous inspection technology and flight path planning for unmanned aerial vehicles on roads [J]. Transportation World, 2025 (35): 8-11
4. [4] Shao Yanhua, Zhang Duo, Chu Hongyu, etc A review of YOLO object detection based on deep learning [J]. Journal of Electronics and Information Technology, 2022, 44 (10): 3697-3708
5. [5] Wang Linyi, Bai Jing, Li Wenjing, etc Research progress on YOLO series object detection algorithms [J]. Computer Engineering and Applications, 2023, 59 (14): 15-29
6. [6] Wang Haiqun, Xie Weimin, Chao Shuai, etc Improved YOLOv11n water surface garbage detection algorithm [J]. Electronic Measurement Technology, 2026, 49 (9): 110-120
7. [7] Yang Yi, Zhang Yuying, Zhao Bin, etc Research on Road Crack Detection Method Based on Improved YOLOv11 [J]. Modern Information Technology, 2026, 10 (2): 84-90
8. [8] Chen Qihao Analysis of Reducing Solar Compressor Unit Faults Based on the "Human Machine Environmental Management" Method [J]. China Equipment Engineering, 2026 (2): 139-142
9. [9] Liu Ziyu, Qian Wenhui, Chen Junxia Research on Risk Factors of Hazardous Chemical Transportation Based on System Dynamics [J]. Hebei Industrial Science and Technology, 2024, 41 (2): 124-132
10. [10] Liu Dianbin, Zhang Qiusheng, Cui Yahui Autonomous inspection method for unmanned aerial vehicles based on machine vision [J]. Intelligent IoT Technology, 2026, 58 (2): 80-85
11. [11] Yao Yue, Yang Xiaohong, Gao Zhiwei, etc Unmanned Aerial Vehicle Small Target Detection Algorithm Based on Improved YOLOv11n [J/OL]. Journal of Zhengzhou University (Science Edition), 2026:1-7 https://doi.org/10.13705/j.issn.1671-6841.2025110.
12. [12] Shen Wei, Zhong Lingshu Overview of Optimization Modeling Methods for Unmanned Aerial Vehicle Inspection [J]. Journal of South China University of Technology (Natural Science Edition), 2026, 54 (6): 12-28
13. [13] Chen Lei, Xing Zhi, Yan Shanpo, etc Improved Small Target Aerial Detection Algorithm Based on YOLOv11n [J]. Electronic Measurement Technology, 2026, 49 (9): 239-248
14. [14] Wang Ning, Zhi Min A review of single-stage universal object detection algorithms under deep learning [J]. Computer Science and Exploration, 2025, 19 (5): 1115-1140

Funding

Jiangsu Province Innovation Capacity Building Plan Project (BM2022013)

About the authors

Li Chong (September 1989-), male, Han ethnicity, from Xuzhou, Jiangsu, senior engineer, master's student, research interests include unmanned and digital safety emergency response.