Hui Luo, Jian Li, Chen Jia. Rail Surface Defect Detection Based on Image Enhancement and Improved Cascade R-CNN[J]. Laser & Optoelectronics Progress, 2021, 58(22): 2212001
- Laser & Optoelectronics Progress
- Vol. 58, Issue 22, 2212001 (2021)
Fig. 1. Overall structure of the method in this paper
Fig. 2. Improved Retinex rail surface image enhancement algorithm framework
Fig. 3. Overall structure diagram of rail surface defect detection algorithm
Fig. 4. RoIAlign diagram
Fig. 5. Examples of rail surface defect image data augmentation
Fig. 6. Changes of training losses based on improved Cascade R-CNN+ResNet-50
Fig. 7. Contrast diagrams of rail surface image enhancement effect
Fig. 8. Comparison diagrams of rail surface defect detection result
|
Table 1. Experimental results of data augmentation effectiveness
|
Table 2. Experimental results of detection performance for different image enhancement algorithms
|
Table 3. Experimental results of different architectures and improved algorithms
|
Table 4. Performance comparison of different algorithms
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