Zhaofeng Huang, Yunchao Tang, Xiangjun Zou, Mingyou Chen, Hao Zhou, Tianlong Zou. Visual Measurement of Crack Width Based on Backbone's Two-Scale Fusion of Features[J]. Laser & Optoelectronics Progress, 2022, 59(10): 1015005
- Laser & Optoelectronics Progress
- Vol. 59, Issue 10, 1015005 (2022)
Fig. 1. Flowchart of the proposed method
Fig. 2. Crack segmentation results of U-net
Fig. 3. Comparison between the classical thinning algorithm and the proposed improved algorithm. (a) Classic thinning algorithm;(b) proposed backbone refining algorithm
Fig. 4. 16 types of neighborhood distribution of backbone points
Fig. 5. Different measurement methods of cracks at the same place
Fig. 6. Judgment of measurement direction using only a single scale feature. (a) Slope feature; (b) neighborhood distribution type feature
Fig. 7. Finding process of measuring points
Fig. 8. Measurable points of a section of crack and its measurement scheme
Fig. 9. Test equipment and environment
Fig. 10. Reduction rate result
Fig. 11. Recall rate of the proposed width measurement direction judgment method
Fig. 13. Field measurement test result
Fig. 14. Deviation between width value and standard value of visual measurement
|
Table 1. Average value and standard deviation of width measurement error
|
Table 2. Statistical results of the recall rate of the direction judgment method
| |||||||||||||||||||||||
Table 3. Direction error statistics of the different methods
|
Table 4. Width measurement error average value and standard deviation
Set citation alerts for the article
Please enter your email address
© Copyright 2018-2021 | Chinese Laser Press. All Rights Reserved 沪ICP备15018463号-20