Abstract:In the realm of special equipment, significant advancements have been achieved in fault detection. Nonetheless, faults originating in the equipment manifest with diverse morphological characteristics and varying scales. Certain faults necessitate the extrapolation from global information owing to their occurrence in localized areas. Simultaneously, the intricacies of the inspection area's background easily interfere with the intelligent detection processes. Hence, a refined YOLOv8 algorithm leveraging the Swin Transformer is proposed, tailored for detecting faults in special equipment. The Swin Transformer serves as the foundational network of the YOLOv8 framework, amplifying its capability to concentrate on comprehensive features during the feature extraction, crucial for fault analysis. A multi-head self-attention mechanism regulated by a sliding window is utilized to expand the observation window's scope. Moreover, an asymptotic feature pyramid network is introduced to augment spatial feature extraction for smaller targets. Within this network architecture, adjacent low-level features are merged, while high-level features are gradually integrated into the fusion process. This prevents loss or degradation of feature information during transmission and interaction, enabling accurate localization of smaller targets. Drawing from wheel–rail faults of lifting equipment as an illustration, the proposed method is employed to diagnose an expanded fault dataset generated through transfer learning. Experimental findings substantiate that the proposed method in adeptly addressing numerous challenges encountered in the intelligent fault detection of special equipment. Moreover, it outperforms mainstream target detection models, achieving real-time detection capabilities.

Video surveillance-based multi-task learning with swin transformer for earthwork activity classification

Post-earthquake Inspection of High-Speed Railway Viaducts with Multi-Scale Task Interaction Deep Learning Strategy

Automatic Recognition of Construction Worker Activities Using Dense Trajectories

Long Video-Based Action Segmentation for Earthmoving Excavators Using Improved Temporal Convolutional Network Models

Automatic excavator action recognition and localisation for untrimmed video using hybrid LSTM-Transformer networks

Earthwork digital twin for teleoperation of an automated bulldozer in edge dumping

UAV Imagery-based Automatic Classification of Ground Surface Types for Earthworks

Vision-Based Action Recognition Of Construction Workers Using Dense Trajectories

StrideNET: Swin Transformer for Terrain Recognition with Dynamic Roughness Extraction

Video-based construction vehicles detection and its application in intelligent monitoring system

End-to-end deep learning for reverse driving trajectory of autonomous bulldozer

Vision-Based Method Integrating Deep Learning Detection for Tracking Multiple Construction Machines

Multi-Swin Mask Transformer for Instance Segmentation of Agricultural Field Extraction

An Improved Method for Detecting Crane Wheel–Rail Faults Based on YOLOv8 and the Swin Transformer

Eswin-Unet: A Collaborative Model for Industrial Surface Defect Detection

A deep learning approach for construction vehicles fill factor estimation and bucket detection in extreme environments

Advanced hybrid LSTM-transformer architecture for real-time multi-task prediction in engineering systems

Integrating Domain Knowledge with Deep Learning Model for Automated Worker Activity Classification in mobile work zone

Multi-Object Detection Method in Construction Machinery Swarm Operations Based on the Improved YOLOv4 Model

Landslide Recognition from Multi-Feature Remote Sensing Data Based on Improved Transformers

A Deep Learning-Based Approach to Enable Action Recognition for Construction Equipment