An elevation measurement method via the fusion of vision and barometric sensors is proposed to achieve convenient and accurate measurement of the elevations of operators during aerial work. A YOLOX deep neural network is constructed to detect the aerial operators and their positions in the pictures of the aerial work site. The changes of real altitude in the relatively short period of climbing are measured via a barometric sensor. Using the barometric elevation measurement and vision detection results, the support vector regression (SVR) is applied to construct and update the regressive model between the image position and the actual elevation of the operator, and the high accuracy elevation measurement results are obtained through the regressive model. In real-world aerial work experiments, the proposed method is compared with the direction elevation measurement using a single barometric sensor and the differential elevation measurement by two barometric sensors. The measurement errors, in the mean absolute error as well as in the root-mean squares-error, of the proposed method are all lower than 0.2 m, and are superior to the two rival methods. The performances of the proposed methods are achieved by fusing the vision detection and the barometric elevation measurement, which overcomes the serious signal shifting and degradation of measurement accuracy of the barometric sensors, and also avoids the laborious calibration procedure of the vision detection system. The proposed method’s elevation measurement accuracy satisfies the requirement on the operator in aerial work, and it is easy to apply in work sites without putting extra burdens on the operators, making the method a practical choice for operator safeguarding in aerial work.
GUOQ. Research on integrated maintenance and detection system of power cables for Internet of Things[D]. Xi’an:Xidian University,2022.(in Chinese)
[3]
肖扬. 电网运维在线监测系统的设计与实现[D]. 成都: 电子科技大学, 2020.
[4]
XIAOY. Design and implementation of online monitoring system for power grid operation and maintenance[D]. Chengdu:University of Electronic Science and Technology of China,2020.(in Chinese)
GUIY G, SUNJ, WUJ, et al .Research on height data fusion algorithm based on GPS/barometric altimeter[J].Machinery & Electronics,2017,35(1):11-14.(in Chinese)
[15]
OVEISA H, GIUSTIE, SGHIOS,et al. A survey on the applications of convolutional neural networks for synthetic aperture radar: Recent advances[J]. IEEE Aerospace and Electronic Systems Magazine,2022, 37(5): 18-42.
ZHANGY T, HUANGD Q, WANGD W, et al. A review on research and application of deep learning-based target detection algorithms[J]. Computer Engineering and Applications, 2023, 59(18): 1-13.(in Chinese)
[18]
ZHANGJ Q, CHENX L, LIY L,et al .Pedestrian detection algorithm based on improved YOLOv3[C]//2021 IEEE International Conference on Power,Intelligent Computing and Systems (ICPICS). July 29-31,2021,Shenyang, China: IEEE,2021: 180-189.
[19]
ZENGT C, WANGJ J, CUIB, et al. The equipment detection and localization of large-scale construction jobsite by far-field construction surveillance video based on improving YOLOv3 and grey wolf optimizer improving extreme learning machine[J]. Construction and Building Materials, 2021, 291: 1-5.
QINX H, HUANGQ D, CHANGD X, et al. Object detection method in open-pit mine based on improved YOLOv5[J]. Journal of Hunan University (Natural Sciences), 2023, 50(2): 23-30.(in Chinese)
[22]
XIAOY Z, TIANZ Q, YUJ C,et al .A review of object detection based on deep learning[J].Multimedia Tools and Applications,2020, 79(33): 23729-23791.
[23]
CHENY C, XUX N .The research of underwater target recognition method based on deep learning[C]//2017 IEEE International Conference on Signal Processing,Communications and Computing (ICSPCC).October 22-25,2017,Xiamen,China: IEEE,2017:1-5.
[24]
王林林 .基于目标检测与动作识别的多目标视觉监管技术研究[D].长春:吉林大学,2021.
[25]
WANGL L. Research on multi-target visual supervision technology based on target detection and motion recognition[D].Changchun: Jilin University, 2021.(in Chinese)
XUK Y. Design of intelligent identification and monitoring system for electric field workers based on image processing technology[D]. Hangzhou: Zhejiang University, 2022. (in Chinese)
[28]
SONGJ, ZHAOY, SONGW, et al. Fisheye image detection of trees using improved YOLOX for tree height estimation[J]. Sensors, 2022, 22(10): 1-6.
[29]
GEZ, LIUS, WANGF, et al. YOLOX: Exceeding YOLO series in 2021[J]. arXiv preprint arXiv:2021.
[30]
ZHANGQ, HUANGY R, SONGR .A ship detection model based on YOLOX with lightweight adaptive channel feature fusion and sparse data augmentation[C]//2022 18th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). November 29 - December 2,2022,Madrid,Spain:IEEE,2022:1-8.
[31]
LIUB Q, HUANGJ Y, LINS,et al .Improved YOLOX-S abnormal condition detection for power transmission line corridors[C]//2021 IEEE 3rd International Conference on Power Data Science (ICPDS).December 26,2021,Harbin,China: IEEE,2021:13-16.
[32]
WANGG, LIUZ, SUNH, et al. YOLOX-BTFPN: An anchor-free conveyor belt damage detector with a biased feature extraction network[J]. Measurement, 2022, 200: 1-6.
[33]
CHENY B, XUP, CHUY Y,et al .Short-term electrical load forecasting using the support vector regression (SVR) model to calculate the demand response baseline for office buildings[J].Applied Energy,2017,195:659-670.
基金资助
中国南方电网有限责任公司创新项目(YNKIXM20210196)
Innovation Program of China Southern Power Grid Co.,Ltd(YNKIXM20210196)