Author Affiliations
1School of Mechanical and Electrical Engineering, Anhui Jianzhu University, Hefei, Anhui 230601, China2Key Laboratory of Construction Machinery Fault Diagnosis and Early Warning Technology, Anhui Jianzhu University, Hefei, Anhui 230601, China3Key Laboratory of Intelligent Manufacturing of Construction Machinery, Anhui Jianzhu University, Hefei, Anhui 230601, Chinashow less
Fig. 1. Infrared polarization information images. (a) 0° azimuth image; (b) 60° azimuth image; (c) 120° azimuth image; (d) s0 image; (e) s1 image; (f) s2 image; (g) DOP image; (h) AOP image
Fig. 2. Process of searching potential target regions. (a) Fusion results of multi azimuth image gray level and information entropy; (b) potential target region images
Fig. 3. Simulation segmentation results of 0°, 60°, 120° azimuth images
Fig. 4. Flow chart of multi-channel infrared polarized thermal image overall segmentation
Fig. 5. Specimen size
Fig. 6. Original images
Fig. 7. Process of multi-channel segmentation with fatigue cycle of 1000 times
Fig. 8. Process of multi-channel segmentation with fatigue cycle of 8000 times
Fig. 9. Process of multi-channel segmentation with fatigue cycle of 14000 times
Fig. 10. Comparison of segmentation results of different algorithms
Parameter | 0° | 60° | 120° |
---|
| 157 | 150 | 142 | H | 380.0 | 331.1 | 336.6 | Emax | 411.2 | 363.4 | 365.2 | ρ | -9.8788×10-5 | -4.2433×10-4 | -8.7336×10-4 |
|
Table 1. Some parameters and calculation results of potential target region at fatigue cycle of 1000 times
Parameter | 0° | 60° | 120° |
---|
| 158 | 147 | 138 | H | 406.8 | 384.5 | 382.5 | Emax | 436.4 | 411.6 | 406.5 | ρ | -2.3927×10-4 | -2.6528×10-4 | -6.1364×10-4 |
|
Table 2. Some parameters and calculation results of potential target region at fatigue cycle of 8000 times
Parameter | 0° | 60° | 120° |
---|
| 154 | 144 | 136 | H | 430.8 | 423.9 | 400.4 | Emax | 433.4 | 447.7 | 422.8 | ρ | -6.5729×10-5 | -1.7110×10-4 | -5.6997×10-4 |
|
Table 3. Some parameters and calculation results of potential target region at fatigue cycle 14000
Algorithm | 1000 times | 8000 times | 14000 times |
---|
FCM | 13.09 | 23.17 | 31.85 | OTSU | 14.82 | 31.57 | 36.07 | MEM | 38.06 | 30.14 | 36.29 | Proposed algorithm | 40.94 | 46.47 | 48.82 |
|
Table 4. GLC of different algorithms unit: %
Algorithm | 1000 times | 8000 times | 14000 times |
---|
FCM | 96.69 | 95.65 | 94.83 | OTSU | 96.34 | 94.17 | 93.66 | MEM | 63.51 | 93.59 | 93.66 | Proposed algorithm | 25.17 | 3.48 | 6.31 |
|
Table 5. δR of different algorithms unit: %