Patent
Rail surface defect detection method based on depth learning network
Hui Zhang,Song Yanan,Li Liu,Hang Zhong,Zhicong Liang +4 more
- 21 Dec 2018
24
TL;DR: Wang et al. as mentioned in this paper presented a rail surface defect detection method based on the depth learning network, aiming at solving various problems existing in the prior-rail detection methods. But the method is limited to the detection and identification of rail surface defects.
read more
Abstract: The invention belongs to the field of depth learning, and provides a rail surface defect detection method based on the depth learning network, aiming at solving various problems existing in the priorrail detection methods. The depth learning method first automatically resets the input rail image to 416*416, and then extracts and processes the image. Image extraction mainly by Darknet-53 model complete. The processing output is mainly accomplished by the FPN-like network model. Firstly, the rail image is divided into cells. According to the position of the defects in the cells, the width, height and coordinates of the center point of the defects are calculated by dimension clustering method, and the coordinates are normalized. At the same time, we use logistic regression to predict the fraction of boundary box object, use binary cross-entropy loss to predict the category contained in the boundary box, calculate the confidence level, and then process the convolution in the output, up-sampling, network feature fusion to get the prediction results. The invention can accurately identify defects and effectively improve the detection and identification rate of rail surface defects.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Patent
A defect target detection method based on an attention mechanism
Zhang Shengsen,Lin Hongzhi,Zheng Zengqiang,Bai Xiang,Liu Ronghua,Shen Yafei +5 more
- 31 May 2019
TL;DR: Wang et al. as mentioned in this paper presented a defect target detection method based on an attention mechanism, which comprises the following steps: marking various defects of all pictures in an original data set to obtain a standard training data set with marks; Obtaining a training label according to the standard training dataset, determining a loss function, obtaining a network model, and performing training by using a reverse conduction method to obtaining a defect regression detection network model.
13
Patent
D-series high-speed train carriage bottom plate bolt fault detection method based on deep learning
He Liwen,Zhang Xingning,Lu Ya +2 more
- 05 Jul 2019
TL;DR: In this paper, a D-series high-speed train carriage bottom plate bolt fault detection method based on deep learning was proposed, which relates to the field of image processing and deep learning.
7
Patent
A pedestrian and vehicle detection method and system based on improved YOLOv3
Liu Tianliang,Wang Guowen,Xie Shipeng,Dai Xiubin +3 more
- 28 May 2019
TL;DR: In this article, a pedestrian and vehicle detection method and system based on improved YOLOv3 is presented, where the cross-layer fusion and reuse of multi-scale features in the backbone network are carried out by adopting a transmittable feature map scale reduction method; and a feature pyramid network is constructed by adopting scale amplification method.
7
Patent
Marine weak and small moving ship real-time detection method based on deep learning
Zhou Weina,Ding Haowen +1 more
- 21 May 2019
TL;DR: In this paper, the authors proposed a real-time detection method based on deep learning for weak and small moving ship detection in a marine monitoring system video, where the network structure is of a full convolutional network type and comprises two parts, and the first part is used for extracting effective features of a target; and the second part was used for detecting a target and determining a final detection result.
4
Patent
Concrete crack identification method based on YOLOv3 deep learning
Shen Yonggang,Yu Zhenwei,Zhang Yiping,Wen Zuolin +3 more
- 27 Aug 2019
TL;DR: In this paper, a multi-target crack recognition method based on a YOLOv3 deep learning algorithm was proposed, which comprises the following steps: importing a crack image into a YOv-3 model, and automatically compressing the image into 416 * 416 pixel resolution; dividing the original image into S * S grids according to the scale size of the feature map.
3
References
•Posted Content
YOLOv3: An Incremental Improvement.
Joseph Redmon,Ali Farhadi +1 more
TL;DR: The authors present some updates to YOLO!
17.8K
Patent
Image crack segmentation method based on deep convolutional neural network
Yao Jian,Liu Yahui,Zhao Jiao,Xie Renping +3 more
- 05 Sep 2017
TL;DR: In this paper, an image crack segmentation method based on a deep convolutional neural network is proposed, which consists of the steps that an original image is inputted to the deep CNN, and a characteristic graph is acquired through convolution, pooling and activation layer characteristic learning; sampling is performed on the characteristic graph so as to obtain the feature graph having the same size with that of the original image; and softmax prediction is performed to obtain a feature graph.
19
Patent
Steel rail surface defect image adaptive segmentation method
Min Yongzhi,Biao Yue,Dang Jianwu,Ma Hongfeng,Zhang Zhenhai,Lin Junting,Zhang Yanpeng,Zhang Xin,Zuo Jing +8 more
- 21 Dec 2016
TL;DR: Wang et al. as mentioned in this paper proposed a steel rail surface defect image adaptive segmentation method based on a row grayscale mean successive summation (RGSM) method.
16
Patent
Method for detecting fissure on surface of subway tunnel
Yu Zujun,Wang Yaodong,Zhu Liqiang,Guo Baoqing,Bai Biao +4 more
- 04 Jun 2014
TL;DR: In this article, a method for detecting a fissure on the surface of a subway tunnel is proposed, in which high-speed image acquisition is conducted by means of a visual system formed by high speed area-array cameras; an acquired image is preprocessed to enable a binary image to be obtained; the communication area of the binary image is calculated, multi-level filtering processing is conducted based on the communication areas, and a tunnel surface image with irregular noise points which are irregularly distributed filtered out is obtained; detection of the fissures on thesurface of the
14
Patent
In-depth learning based image recognition method
Xingpei Xu,Song Yuqing,Hu Lu +2 more
- 13 Apr 2018
TL;DR: In this paper, an in-depth learning based image recognition method was proposed, which belongs to the technical field of machine learning and can effectively judge the type of the input medical input according to different classification standards; at the same time, the calculation cost can be reduced substantially and assistance can be provided to a doctor in disease diagnosis.
12