Liangfu Li, Nan Wang, Biao Wu, Xi Zhang. Segmentation Algorithm of Bridge Crack Image Based on Modified PSPNet[J]. Laser & Optoelectronics Progress, 2021, 58(22): 2210001
- Laser & Optoelectronics Progress
- Vol. 58, Issue 22, 2210001 (2021)
Fig. 1. Low-quality image preprocessing. (a) Original image; (b) result obtained by Laplacian sharpening
Fig. 2. Schematic of bridge crack dataset amplification. (a) Original image; (b) rotate 180°; (c) rotate 270°; (d) horizontal flip; (e) vertical flip
Fig. 3. PSPNet structure
Fig. 4. Proposed network structure
Fig. 5. Schematic of the 3×3 convolution kernel receptive field with expansion rates of 1, 2, 3
Fig. 6. SPAM structure
Fig. 7. Loss curve during training
Fig. 8. Comparison results without and with SPAM. (a) Original image; (b) label; (c) without SPAM; (d) with SPAM
Fig. 9. Comparison of results between different algorithms. (a) Original image; (b) label; (c) Deeplab-V3+; (d) U-Net++; (e) EncNet; (f) proposed algorithm
|
Table 1. Comparison of experiment results without and with SPAM
|
Table 2. Comparison between the proposed algorithm and mainstream semantic segmentation models
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