Fault Detection of Train Height Valve Based on Nanodet-Resnet101

Zhou, Zhonghe; Hu, Yuanjiang; Deng, Xue; Huang, Deqing; Yi, Lin

doi:10.1109/yac53711.2021.9486540

Cited by 7 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Zhang et al [10] proposed a fault detection method to detect rod springs of fixtures with an experimental accuracy of 91.98%. Zhou et al [11] proposed a detection algorithm for height valve faults, which could detect faults with an accuracy of 97%. Ye et al [12] proposed a detector used the K-means algorithm to design the anchors.…”

Section: A Fault Detection Of Freight Train Imagesmentioning

confidence: 99%

Visual Fault Detection of Multiscale Key Components in Freight Trains

Zhang

Zhou

Pan

et al. 2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

Fault detection for key components in the braking system of freight trains is critical for ensuring railway transportation safety. Despite the frequently employed methods based on deep learning, these fault detectors are extremely reliant on hardware resources and complex to implement. In addition, no train fault detectors consider the drop in accuracy induced by scale variation of fault parts. This paper proposes a lightweight anchor-free framework to solve the above problems. Specifically, to reduce the amount of computation and model size, we introduce a lightweight backbone and adopt an anchor-free method for localization and regression. To improve detection accuracy for multi-scale parts, we design a feature pyramid network to generate rectangular layers of different sizes to map parts with similar aspect ratios. Experiments on four fault datasets show that our framework achieves 98.44% accuracy while the model size is only 22.5 MB, outperforming state-of-the-art detectors.

show abstract

Section: A Fault Detection Of Freight Train Imagesmentioning

confidence: 99%

Visual Fault Detection of Multiscale Key Components in Freight Trains

Zhang

Zhou

Pan

et al. 2023

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

show abstract

“…; however, the classification network model parameters and computation are large, which improves the detection accuracy to a certain extent, but it is difficult to deploy the network model to mobile devices with limited resources, so in order to meet special scenarios, the classification network needs to be lightweight. There are two mainstream ways of lightweighting: one is lightweight basic networks, such as SqueezeNet [3], Xception [4], MobileNet series [5], ShuffleNet series [8], GhostNet [10], etc., which reduce the number of model parameters by using depth-separable convolution, adjustable hyperparameters, etc.,more highly compact lightweight networks such as YOLONano [11], NanoDet [12], etc. The other is to compress the overall network parameters, and there are mainly methods of model compression such as network weight pruning [13], quantization [14], and knowledge distillation [15], etc.…”

Section: Introductionmentioning

confidence: 99%

Lightweight YOLOx-based transmission line insulators and their defects detection

Lu,

Ruan,

Wang

et al. 2024

Fifteenth International Conference on Graphics and Image Processing (ICGIP 2023)

View full text Add to dashboard Cite

A lightweight network model based on YOLOx is proposed for the problems of limited resources of transmission line UAV inspection platform, high complexity of target detection algorithm and slow inference speed. First, the lightweight ShuffleNetV2_Plus network is used as the backbone network for feature extraction, and the Depthwise Convolution (DWConv) in the ShuffleNetV2 network is expanded by replacing 3×3DWConv in the ShuffleUnit module with 5×5DWConv in the ShuffleUnit module, and prune the convolution layer of the model, and prune the 1×1Pointwise Convolution (PWConv) in the ShuffleUnit basic unit module to reduce the network parameters while increasing the network perceptual field. At the same time, add the Efficient Channel Attention (ECA) module in the neck feature fusion part to make the network better focus on important regions and improve the target detection accuracy at a small computational cost. Finally, the ordinary convolution in the YOLOx detection decoupling head is replaced with Depthwise Separable Convolution (DSConv) to further reduce the model complexity. The results show that the inference time of the lightweight network model proposed in this paper is only 5.8ms, the model parameters are only 4.361MB, and the FLOPs are only 10.725G, and the detection accuracy is high on the combined self-built transmission line dataset.

show abstract