Crack Detection Algorithm Based on Improved Multibranch Feature Shared Structure Network

Gang Li; Yongqiang Chen; Tingquan He; Yu Dai; Dongchao Lan

doi:10.3788/LOP202259.1215005

Laser & Optoelectronics Progress, Volume. 59, Issue 12, 1215005(2022)

Crack Detection Algorithm Based on Improved Multibranch Feature Shared Structure Network

Gang Li¹, Yongqiang Chen^1、*, Tingquan He², Yu Dai¹, and Dongchao Lan¹

Author Affiliations

¹School of Electronics and Control Engineering, Chang’an University, Xi’an 710064, Shaanxi , China

²Information Department, Guangxi New Development Transportation Group Co., Ltd., Nanning 530029, Guangxi , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(16)

Fig. 1. Depth separable convolution. (a) Convolution process; (b) convolution structure

Download full size

Fig. 2. Pyramid pooling module

Download full size

Fig. 3. Improved multi-branch feature shared structure network

Download full size

Fig. 4. Structure diagram of modules. (a) GCN module; (b) BR module

Download full size

Fig. 5. Structure diagram of RRC module

Download full size

Fig. 6. Process of dataset annotation and crop

Download full size

Fig. 7. Training loss and accuracy obtained by changing the initial learning rate. (a) Loss; (b) accuracy

Download full size

Fig. 8. Visual comparison results of different algorithms in crack images with noise

Download full size

Fig. 9. Change trend of each indicator. (a) Evaluation index; (b) PR curve

Download full size

Fig. 10. Segmentation effects of the proposed algorithm on the public dataset

Download full size

Fig. 11. Schematic of visualized results of crack skeleton extraction

Download full size

Fig. 12. Comparison of measurement errors of crack length and width

Download full size

Table 1. Partition of crack dataset
View table
Table 1. Partition of crack dataset
Parameter Training set Validation set Test set
Image size /（pixel×pixel） $512 \times 512$ $512 \times 512$ $512 \times 512$
Number of images /frame 12000 4000 4000

Table 2. Performance comparison of different crack detection algorithms

View table

Table 2. Performance comparison of different crack detection algorithms

Algorithm	Precision	Recall	F1	MIoU	FPS（Millisecond/image）
Fast-SCNN	0.7548	0.7962	0.7749	0.7692	67.5（14.81）
SegNet	0.8312	0.7864	0.8082	0.8206	18.7（53.48）
Pspnet	0.9021	0.9115	0.9068	0.8021	29.3（34.13）
DeepCrack	0.8463	0.8159	0.8308	0.8642	25.6（39.06）
CrackU-Net	0.8736	0.7938	0.8318	0.8527	34.8（28.74）
Proposed algorithm	0.9695	0.9283	0.9485	0.8902	59.4（16.84）

Table 3. Test results of the proposed algorithm on different public datasets
View table
Table 3. Test results of the proposed algorithm on different public datasets
Dataset Precision Recall F1 MIoU FPS（Millisecond/image）
Cracktree200 0.9432 0.9509 0.9470 0.8772 54.5（18.35）
GAPs384 0.9527 0.9445 0.9486 0.8495 53.8（18.59）
Crack500 0.9683 0.9485 0.9583 0.8679 56.2（17.79）

Table 4. Relative errors of length and width of different comparison algorithms
View table
Table 4. Relative errors of length and width of different comparison algorithms
Parameter Fast-SCNN SegNet Pspnet DeepCrack CrackU-Net Proposed algorithm
Relative error of length /% 8.62 7.28 6.96 5.73 5.14 4.73
Relative error of width /% 7.94 6.53 7.32 6.87 6.02 5.21

Tools

Get Citation

Copy Citation Text

Gang Li, Yongqiang Chen, Tingquan He, Yu Dai, Dongchao Lan. Crack Detection Algorithm Based on Improved Multibranch Feature Shared Structure Network[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1215005

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites