DPIM algorithm for hoisting operation scene based on inspection robot

doi:10.16265/j.cnki.issn1003-3033.2024.07.0235

Abstract

Abstract:

In order to improve the high-precision detection and early warning of crane safety operation and enhance the safety management ability of enterprises, focusing on the needs of unmanned industrial safety incident analysis and monitoring and early warning, an inspection robot that combines ground and air flight in hoisting scene was customized to intelligentize hoisting safety monitoring, pop-up image recording and safety alarm.A lifting dataset Cranes-Dataset (CRN-Dataset) containing 3 120 images was made, and DPIM algorithm was proposed to enhance the rapid detection ability of multi-scale objects.Based on corner detection and density-based spatial clustering of applications with noise and considering the safety attributes of the space distance between cranes and workers, the process of triggering alarms based on safety rules was developed to record real-time illegal operation image and popup alarm.The results show that, after actual deployment and verification, the DPIM algorithm significantly improves target identification ability compared with other traditional algorithms, and it is suitable for real-time calculation and data transmission of embedded edge intelligent analysis nodes to complete field deployment.

Key words: inspection robot, hoisting scene, dynamic perspective intelligent monitoring (DPIM) algorithm, personnel denial, edge intelligent analysis

CLC Number:

X948

LIN Shikang, HOU Qingwen, GUAN Yuyin, WANG Wencai, LI Jialu, CHEN Xianzhong. DPIM algorithm for hoisting operation scene based on inspection robot[J]. China Safety Science Journal, 2024, 34(7): 146-152.

Figures/Tables 11

Fig.1

Table 1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Table 2

Table 3

References 15

[1]	GU Yuwan, XU Shoukun, WANG Yaru, et al. An advanced deep learning approach for safety helmet wearing detection[C]. International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData), 2019: 669-674.
[2]	DUAN Qipeng, PING Kuang, LI Fan, et al. Method of safety helmet wearing detection based on key-point estimation without anchor[C]. The 17^th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP), 2020: 93-96.
[3]	DENG Benyang, LEI Xiaochun, YE Miao. Safety helmet detection method based on YOLOv4[C]. The 16^th International Conference on Computational Intelligence and Security (CIS), 2020: 155-158.
[4]	GAO Guoqing, LIN Wenhua, LI Guohe, et al. Identification of risk behavior in power working scenarios based on deep learning[C]. The 10^th Joint International Information Technology and Artificial Intelligence Conference (ITAIC), 2022: 391-395.
[5]	赵江平, 刘星星, 张想卓. 基于改进YOLOv5s的外脚手架隐患图像识别技术[J]. 中国安全科学学报, 2023, 33(12): 60-66. doi: 10.16265/j.cnki.issn1003-3033.2023.12.2011
	ZHAO Jiangping, LIU Xingxing, ZHANG Xiangzhuo. Research on image recognition technology for external scaffold hidden danger based on improved YOLOv5s[J]. China Safety Science Journal, 2023, 33(12): 60-66. doi: 10.16265/j.cnki.issn1003-3033.2023.12.2011
[6]	LI Zhishan, XIE Wenqing, ZHANG Lingzhi, et al. Toward efficient safety helmet detection based on YOLOv5 with hierarchical positive sample selection and box density filtering[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-14.
[7]	SUN Weichen, YANG Zhanhua, ZHAO Bo, et al. Research on target detection of regional monitoring with complex background using CNN and background modelling[C]. The 7^th International Conference on Machine Vision and Information Technology (CMVIT), 2023:102-106.
[8]	FANG Lijin, HUANG Yongjie, ZHAO Qiankun, et al. Monocular 3D object detection based on occlusion optimization[C]. The 7^th International Conference on Intelligent Computing and Signal Processing (ICSP), 2022:604-609.
[9]	INTAWONG K, PROMPANTAKORN P, BOONCHIENG W. An improved automatic zone detection algorithm based on mechanic flow for complex scene event[C]. Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunication Engineering, 2021:265-268.
[10]	HEO J, WANG Y, PARK J. Occlusion data augmentation for object detectors based on random erasing and bounding box tailoring[C]. The 20^th International Conference on Control, Automation and Systems (ICCAS), 2020: 1068-1072.
[11]	GALLO G, FRANCESCO D R, GARZELLI F, et al. A smart system for personal protective equipment detection in industrial environments based on deep learning at the edge[J]. IEEE Access, 2022, 10: 110 862-110 878.
[12]	张磊, 李熙尉, 燕倩如, 等. 基于改进YOLOv5s的综采工作面人员检测算法[J]. 中国安全科学学报, 2023, 33(7): 82-89. doi: 10.16265/j.cnki.issn1003-3033.2023.07.2226
	ZHANG Lei, LI Xiwei, YAN Qianru, et al. Personnel detection algorithm in fully mechanized coal face based on improved YOLOv5s[J]. China Safety Science Journal, 2023, 33(7): 82-89. doi: 10.16265/j.cnki.issn1003-3033.2023.07.2226
[13]	HE Kaiming, CHEN Xinlei, XIE Saining, et al. Masked autoencoders are scalable vision learners[C]. The IEEE Conference on Computer Vision and Pattern Recognition, 2022:16000-16 009.
[14]	KAT R, JEVNISEK R, AVIDAN S. Matching pixels using co-occurrence statistics[C]. The IEEE Conference on Computer Vision and Pattern Recognition, 2018: 1751-1759.
[15]	LINDENBERGER P, SARLIN P, POLLEFEYS M. LightGlue: local feature matching at light speed[C]. International Conference on Computer Vision (ICCV), 2023: 17 581-17 592.

CRN-Dataset	图像	吊车	工人
训练集	2 184	2 380	2 795
验证集	624	665	802
测试集	312	333	400
总和	3 120	3 378	3 997

模型	P	R	mAP@0.5	FPS/(帧·s^-1)
YOLOv5l	0.862	0.771	0.778	47.6
Faster R-CNN	0.704	0.417	0.655	35.5
SSD	0.440	0.275	0.357	52.9
YOLOv8s	0.838	0.696	0.701	41.2
DPIM	0.894	0.809	0.831	39.4

模型	P	R	mAP@0.5	FPS/(帧·s^-1)
YOLOv5l	0.862	0.771	0.778	47.6
YOLOv5l+C2f	0.870	0.777	0.791	50.3
YOLOv5l+CA	0.882	0.790	0.815	36.9
YOLOv5l+C2f+CA	0.892	0.807	0.820	39.8