Transmission line inspection is an important means to ensure the stable transportation of electricity, and existing hidden danger detection methods are difficult to effectively model global information while ensuring real-time model performance. This article introduced a state space model based on YOLOv8, and achieved long-distance hidden danger detection in space through block operation and low rank approximation, and global information was extracted with lower complexity. In the state space model, cross scanning and dynamic multi-path activation mechanisms were used to address the direction insensitivity and non-causal characteristics of the state space model to image data, capture hidden target structure and pattern information, and using spatial context information to identify local features; In the stage of target classification and localization, a separated detection head based on spatial alignment was designed to align spatial misalignment through different representation methods to enhance the accuracy of classification and localization. Finally, experiments show that the new model outperforms mainstream single-stage and two-stage object detection models in terms of average accuracy and frame rate with an accuracy improvement of 2.8%.
LIUChuanyang, WUYiquan.Research progress of vision detection methods based on deep learning for transmission lines[J].Proceedings of the CESS, 2023, 43(19): 7423-7446.(in Chinese)
LIUChuanyang, WUYiquan, LIUJingjing.Research progress of vision-based rust defect detection methods for metal fittings in transmission lines[J].Chinese Journal of Scientific Instrument, 2024, 45(3): 286-305.(in Chinese)
LUJiazheng, ZHOUTejun, WUChuanping, et al.Fault statistics and analysis of 220 kV and above power transmission line in province-level power grid[J].High Voltage Engineering, 2016, 42(1): 200-207.(in Chinese)
PUTianjiao, QIAOJi, HANXiao, et al.Research and application of artificial intelligence in operation and maintenance for power equipment[J].High Voltage Engineering, 2020, 46(2): 369-383.(in Chinese)
LINYifeng, WANGZengping, WANGTong, et al.Discussion on building a stronger first line of defense for grid security[J]. Power System Protection and Control, 2022, 50(16): 170-178. (in Chinese)
ZHAIYongjie, YANGKe, WANGQianming, et al.Disc insulator defect detection based on mixed sample transfer learning[J].Proceedings of the CSEE, 2023, 43(7): 2867-2877. (in Chinese)
LIBin, QULuyao, ZHUXinshan, et al. Insulator defect detection based on multi-scale feature fusion[J].Transactions of China Electrotechnical Society, 2023, 38(1): 60-70.(in Chinese)
ZANGJiye, QUZhaoyang, DONGYunchang, et al.Target detection model of cascade distillation transmission line based on bionic visual backbone[J].High Voltage Engineering, 2024, 50(8): 3757-3768. (in Chinese)
[21]
LIK C, LIX H, WANGY, et al.VideoMamba: State space model for efficient video understanding[DB/OL]. (2024-03-12)[2024-07-07].https:arxiv.org/abs/2403.06977.
[22]
GUA, DAO T. Mamba: Linear-time sequence modeling with selective state spaces[DB/OL].(2024-05-31)[2024-07-07].
[23]
LIUY, TIANY, ZHAOY, et al. Vmamba: Visual state space model[DB/OL].(2024-01-18)[2024-07-07].
[24]
KHANS, NASEERM, HAYATM, et al. Transformers in vision: A survey[DB/OL].(2022-01-19)[2024-07-07].
[25]
赵丽娟.输电线路组件及其缺陷检测的关键技术研究[D].北京: 华北电力大学, 2023.
[26]
ZHUX, SUW, LUL, et al. Deformable DETR: Deformable transformers for end-to-end object detection[DB/OL].(2021-03-18)[2024-07-07].
[27]
LIUW, ANGUELOVD, ERHAND, et al. SSD: Single shot multibox detector[C]//Computer Vision-ECCV, 2016: 21-37.
[28]
VU T, JANGH, PHAMT X, et al.Cascade RPN: Delving into high-quality region proposal network with adaptive convolution[DB/OL].(2019-12-04)[2024-07-07].