本文選題：塞釘 + 軌道電路　； 參考：《中國(guó)鐵道科學(xué)》2017年03期

【摘要】：根據(jù)高鐵巡檢車(chē)所采集軌腰圖像中鐵道塞釘圖像的特點(diǎn),在既有計(jì)算機(jī)視覺(jué)的目標(biāo)檢測(cè)算法的基礎(chǔ)上,提出基于深度學(xué)習(xí)的鐵道塞釘自動(dòng)檢測(cè)算法。在目標(biāo)檢測(cè)的區(qū)域選擇階段,借鑒顯著性檢測(cè)的思路,提出余譜區(qū)域候選(Spectrum Residual Region Proposal,SRP)算法,即利用含塞釘?shù)能壯鼒D像與不含塞釘?shù)能壯骄鶊D像之間的頻譜差異,通過(guò)快速傅里葉變換,得到兩圖像間的幅度譜差的絕對(duì)值(余譜),再通過(guò)快速傅里葉反變換及后處理,得到候選目標(biāo)區(qū)域;然后在目標(biāo)檢測(cè)的特征提取階段,設(shè)計(jì)塞釘卷積神經(jīng)網(wǎng)絡(luò)(plug Convolution Neural Network,pCNN),該網(wǎng)絡(luò)通過(guò)4個(gè)卷積層、3個(gè)池化層、3個(gè)非線性變換層、3個(gè)規(guī)范化層、2個(gè)全連接層和1個(gè)泄露層,自動(dòng)從候選目標(biāo)區(qū)域逐層提取最能表現(xiàn)塞釘特征的特征圖像;最后基于特征圖像采用支持向量機(jī)(SVM)的分類(lèi)器判斷候選目標(biāo)區(qū)域是否含有塞釘,從而實(shí)現(xiàn)塞釘?shù)淖詣?dòng)定位。大量實(shí)際測(cè)試以及與其他算法比較的結(jié)果表明,該算法的檢測(cè)效果最優(yōu)。
[Abstract]:According to the characteristics of railway stud images collected by high-speed railway inspection vehicle, an automatic detection algorithm based on depth learning is proposed on the basis of the existing target detection algorithms of computer vision. In the region selection stage of target detection, using the idea of significant detection for reference, this paper proposes a candidate Spectrum Residual Region Proposal Residual Region algorithm for cospectral region, that is, using the spectral difference between the rail waist image with studs and the average rail waist image without studs. The absolute value of amplitude spectral difference between two images (cospectrum) is obtained by fast Fourier transform (FFT), and then the candidate target region is obtained by FFT and post-processing, and then in the feature extraction stage of target detection, Plug Convolution Neural network pCNNs are designed. The network consists of four convolution layers, three pool layers, three nonlinear transformation layers, three normalized layers, two fully connected layers and one leak layer. Finally, the feature image is extracted from the candidate target area layer by layer. Finally, the support vector machine (SVM) classifier is used to determine whether the candidate target region contains studs or not, so as to realize the automatic location of studs. A large number of practical tests and comparison with other algorithms show that the algorithm has the best detection effect.
【作者單位】： 中國(guó)鐵道科學(xué)研究院基礎(chǔ)設(shè)施檢測(cè)研究所;
【基金】：國(guó)家“九七三”計(jì)劃項(xiàng)目(2013CB329400) 中國(guó)鐵路總公司科技研究開(kāi)發(fā)計(jì)劃重大項(xiàng)目(2015T003-A) 中國(guó)鐵道科學(xué)研究院行業(yè)服務(wù)技術(shù)創(chuàng)新項(xiàng)目(2014YJ052)
【分類(lèi)號(hào)】：TP18;U216.3
，

本文編號(hào)：1887332