Defect Detection Model for Catenary Components in Environments with Few Samples and Small Targets
Journal Title: Urban Mass Transit - Year 2024, Vol 27, Issue 12
Abstract
［Objective］ Aiming at the difficulties of detecting small-volume parts of rail transit catenary under few-sample conditions, a defect detection method integrating generative adversarial networks and deep segmentation models is proposed. ［Method］ The composition of the detection system, the improved DCGAN (deep convolutional generative adversarial network) structure and the improved YOLACT (single-stage conditional adversarial network) model are introduced, and the defect detection effect is verified based on the actual defect datasets. ［Result & Conclusion］ In the sample expansion part, the improved DCGAN model can improve the quality of generated samples by introducing several normalization operations and adding an efficient channel attention mechanism. In the model detection part, the improved ResNeXt-FPN (deep con-volutional neural network combining residual network and feature pyramid network) structure is used to replace the backbone network of the original YOLACT model, aiming to fully characterize the multi-scale features of the target. The introduction of CS (channel and space) attention mechanism in the mask branch network can effectively improve the detection accuracy of small-volume parts. The proposed defect detection model can achieve high-precision detection of cotter pins in complex contact network images, with a detection accuracy and recall rate of up to 88.63% and 87.49% respectively. Compared with the original YOLACT model, the comprehensive performance of the proposed defect detection model is improved by about 6.2%.
Authors and Affiliations
Xiaosi YU, Baoquan WEI, Zewen LI, Fangming DENG
Keywords
Related Articles
- EP ID EP752720
- DOI 10.16037/j.1007-869x.2024.12.015
- Views 17
- Downloads 0
Passing Performance of Inverted T-shaped Frog Turnouts for Suspended Monorail Vehicles
[Objective] The suspended monorail transportation system holds promising prospects, yet researches on the dynamics performance of vehicle passing through turnouts remain relatively scarce, requiring comparison for furthe...
Identification Algorithm of Rail Transit Pantograph-catenary Contact Force Abnormality Based on LSTM Prediction Error
Objective Contact force is a crucial aspect of comprehensive detection in the rail transit PC (pantograph-catenary) system and serves as an important evaluation factor of PC system performance. However, during the detect...
Construction Technology for Ultra-wide Excavation near Existing Buildings in Metro Station Shield Tunneling Receiving Area
Objective It is aimed to address the difficulty in settlement control of existing ground buildings during ultra-large section excavation beneath them in metro station construction. Method Based on the shield tunneling re...
Metro Train Passenger Evacuation Model Based on NFPA 130:2023 Standard for Fixed Guideway Transit and Passenger RailSystems
［Objective］ Due to the relatively confined space and high passenger capacity of metro trains, fire safety and life protection are attracting significant attentions in the public transportation industry. To evaluate the...
Research on Site Selection of Emergency Rescue Station in Urban Rail Transit Line Network Based on P-center Model
Objective With the expansion of urban rail transit line network, the system risks become more complicated and uncontrollable. In order to improve the emergency rescue capability of urban rail transit system, site selecti...