Catenary dropper fault identification based on improved FCOS algorithm

Abstract

The contact network dropper works in a harsh environment, and suffers from the impact effect of pantographs during running of trains, which may lead to faults such as slack and broken of the dropper wire and broken of the current-carrying ring. Due to the low intelligence and poor accuracy of the dropper fault detection network, an improved fully convolutional one-stage (FCOS) object detection network was proposed to improve the detection capability of the dropper condition. Firstly, by adjusting the parameter α in the network focus loss function, the problem of positive and negative sample imbalance in the network training process was eliminated. Secondly, the generalized intersection over union (GIoU) calculation was introduced to enhance the network’s ability to recognize the relative spatial positions of the prediction box and the bounding box during the regression calculation. Finally, the improved network was used to detect the status of dropper pictures. The detection speed was 150 sheets per millisecond, and the MAP of different status detection was 0.951 2. Through the simulation comparison with other object detection networks, it was proved that the improved FCOS network had advantages in both detection time and accuracy, and could identify the state of dropper accurately.

Keywords

catenary dropper / fully convolutional one-stage (FCOS) network / defect identification / generalized intersection over union (GIoU) / focal loss

References

[1] MEI X Y, GU G M, CHEN C. State identification of catenary clamp based on deep learning. Journal of Lanzhou Jiaotong University, 2022, 41(1): 61-67.

[2] LUO F L, YE W, WANG J. Defect detection of the puller bolt in high-speed railway catenary based on deep learning. Journal of Railway Science and Engineering, 2021, 18(3): 605-614.

[3] YU G W. Application of image processing based on hanging string defect detection algorithm in 4C system. Electrified Railway, 2021, 32(S1): 154-158.

[4] HUANG Y M, YUAN T C, YANG J. Research on identification method of catenary fault based on SVM. Computer Integrated Manufacturing Systems, 2018, 35(11): 145-152.

[5] HAN Y, LIU Z G, GENG X, et al. Feature detection of ear pieces in catenary support devices of high-speed railway based on HOG features and two-dimensional gabor transform. Journal of the China Railway Society, 2017, 39(2): 52-57.

[6] BAI R M. Detection and recognition of catenary string and pantograph skateboard based on image processing technology. Chengdu: Southwest Jiaotong University, 2017.

[7] YU X N, GU G M, WANG Y P, et al. Catenary dropper fault detection method based on faster R-CNN. Journal of Lanzhou Jiaotong University, 2021, 40(2): 58-65.

[8] HU D, JIN W D, TANG P. Detection of railway dropper based on double scale features. Electric Engineering, 2020, 15(20): 64-68.

[9] BIAN J P, HAO J X, ZHAO S, et al. Fault identification and location of catenary suspension based on improved capsule network. Transactions of China Electrotechnical Society, 2020, 35(24): 5187-5196.

[10] CHEN Q, PENG J S, YAN Y F, et al. Method based on FCOS and resNet50-FL for identifying stress-free Dropper. Journal of the China Railway Society, 2021, 43(10): 36-42.

[11] WANG Q H, GU W, CAI P Z, et al. Detection method of double side breakage of population cotton seed based on improved YOLO v4. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(1): 389-397.

[12] HOU Q Z, SUN J Y, WANG H, et al. Runway edge lights brightness detection based on improved RetinaNet. Laser & Optoelectronics Progress, 2022, 59(2): 192-200.

[13] YIN T P, YANG J. Detection of steel surface defect based on faster R-CNN and FPN//2021 7th International Conference on Computing and Artificial Intelligence, April 23-26, 2021, Tianjin, China. New York: ACM, 2021: 15-20.

[14] GU G M, CHEN C, YU X N, et al. Target location algorithm of contact network pipe cap based on improved Faster R-CNN. Laser & Optoelectronics Progress, 2022, 59(4): 140-150.

[15] ZHANG Z H, JIA W K, SHAO W J, et al. Green apple based on optimized FCOS in orchards. Spectroscopy and Spectral Analysis, 2022, 42(2): 647-653.

[16] TAN Y D, YU D, HU Y. An application of an improved FCOS algorithm in detection and recognition of industrial instruments. Procedia Computer Science, 2021, 183: 237-244.

[17] LIU S, CHI J N, WU C D. FCOS-Lite: an efficient anchor-free network for real-time object detection//2021 33rd Chinese Control and Decision Conference (CCDC), May 22-24, 2021, Kunming, China. New York: IEEE, 2021: 1519-1524.

[18] LONG Y, LI N N, GAO Y, et al. Apple fruit detection under natural condition using improved FCOS network. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(12): 307-313.

[19] TIAN Z, SHEN C H, CHEN H, et al. FCOS: fully convolutional one-stage object detection//2019 IEEE/CVF International Conference on Computer Vision (ICCV), October 27-November 2, 2019, Seoul, Korea (South). New York: IEEE, 2019: 9626-9635.

[20] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(2): 318-327.

[21] REZATOFIGHI H, TSOI N, GWAK J, et al. Generalized intersection over union: a metric and a loss for bounding box regression//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 15-20, 2019, Long Beach, CA, USA. New York: IEEE, 2019: 658-666.