Semi-Supervized Crack-Detection Method Based on Image-Semantic Segmentation

Liu Pei; Huang Yaping

doi:10.3788/LOP202158.0615004

Laser & Optoelectronics Progress, Volume. 58, Issue 6, 615004(2021)

Semi-Supervized Crack-Detection Method Based on Image-Semantic Segmentation

Liu Pei^1,2 and Huang Yaping^1,2、*

¹Beijing Key Lab of Traffic Data Analysis and Mining, School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China

²School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(13)

Fig. 1. Semi-supervised training framework

Download full size

Fig. 2. MSCM structure

Download full size

Fig. 3. Improved network framework

Download full size

Fig. 4. Pseudo tags obtained by different methods. (a) Original images; (b) ground-truth; (c) SF method; (d) wCtr method; (e) GC method

Download full size

Fig. 5. Crack-detection effect under different networks. (a) Original images; (b) ground-truth; (c) SegNet network; (d) DeepCrack network; (e) proposed network

Download full size

Fig. 6. Cack detection effect after training with different proportion of manual tag and pseudo tag. (a) Original images; (b) ground-truth; (c) 0; (d) 1/150; (e) 1/65; (f) 1/30; (g) 1/15; (h) 1/6; (i) 1

Download full size

Fig. 7. Detection effect of different cracks in different networks under full supervision. (a) Original images; (b) ground-truth; (c) SegNet network; (d) DeepCrack network; (e) proposed network

Download full size

Table 1. Crack-detection results of different networks
View table
Table 1. Crack-detection results of different networks
Network Precision /% Recall /% F1-score
SegNet^[19] 85.23 89.77 86.74
DeepCrack^[14] 81.71 92.88 86.26
Proposed network 80.29 93.56 85.66

Table 2. Quantitative analysis results of three methods
View table
Table 2. Quantitative analysis results of three methods
Method Precision /% Recall /% F1-score
SF^[16] 76.87 72.37 69.22
wCtr^[17] 74.33 71.30 67.34
GC^[18] 86.57 66.93 70.00

Table 3. Quantitative analysis results of manual labeled dataset and pseudo labeled dataset in different proportions
View table
Table 3. Quantitative analysis results of manual labeled dataset and pseudo labeled dataset in different proportions
Proportion Precision /% Recall /% F1-score
0 72.93 85.21 74.52
1/150 72.74 88.62 76.59
1/65 77.48 88.50 80.47
1/30 77.68 92.67 83.39
1/15 80.29 93.56 85.66
1/6 85.99 92.17 88.42
1 89.34 97.48 92.31

Table 4. Quantitative analysis results after pre-training on different proportions of manually annotated datasets
View table
Table 4. Quantitative analysis results after pre-training on different proportions of manually annotated datasets
Proportion Precision /% Recall /% F1-score
1/150 14.26 50.35 21.46
1/65 25.85 60.38 35.04
1/30 49.83 63.80 54.81
1/15 69.48 83.51 75.21
1/6 82.16 91.59 86.22

Table 5. Crack detection results of SF, wCtr and GC methods under different fusion coefficients
View table
Table 5. Crack detection results of SF, wCtr and GC methods under different fusion coefficients
Fusion coefficient Precision /% Recall /% F1-score
1∶0∶0 81.78 89.81 84.81
0∶1∶0 81.07 88.14 83.68
0∶0∶1 85.32 95.03 89.34
1∶1∶1 80.29 93.56 85.66
1∶1∶2 85.93 91.33 87.69
Union 80.39 91.67 84.77

Table 6. Crack detection results of different networks under full supervision
View table
Table 6. Crack detection results of different networks under full supervision
Network Precision /% Recall /% F1-score
SegNet ^[19] 85.23 89.77 86.74
DeepCrack^[14] 81.71 92.88 86.26
Proposed network 89.34 96.48 93.31

Tools

Get Citation

Copy Citation Text

Liu Pei, Huang Yaping. Semi-Supervized Crack-Detection Method Based on Image-Semantic Segmentation[J]. Laser & Optoelectronics Progress, 2021, 58(6): 615004

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Machine Vision

Received: Jul. 10, 2020

Accepted: --

Published Online: Mar. 11, 2021

The Author Email: Yaping Huang (yphuang@bjtu.edu.cn)

DOI:10.3788/LOP202158.0615004

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Crack-detection results of different networks

Table 1. Crack-detection results of different networks

Table 2. Quantitative analysis results of three methods

Table 2. Quantitative analysis results of three methods

Table 3. Quantitative analysis results of manual labeled dataset and pseudo labeled dataset in different proportions

Table 3. Quantitative analysis results of manual labeled dataset and pseudo labeled dataset in different proportions

Table 4. Quantitative analysis results after pre-training on different proportions of manually annotated datasets

Table 4. Quantitative analysis results after pre-training on different proportions of manually annotated datasets

Table 5. Crack detection results of SF, wCtr and GC methods under different fusion coefficients

Table 5. Crack detection results of SF, wCtr and GC methods under different fusion coefficients

Table 6. Crack detection results of different networks under full supervision

Table 6. Crack detection results of different networks under full supervision