Figure 1.
The basic information of the Radarsat-2 fully polarimetric synthetic aperture radar (PolSAR) data in the Yushu study site: (a) Google Earth image, showing the coverage of the PolSAR data (red rectangle) and the location of Yushu County (yellow rectangle); (b) the Pauli RGB image of the PolSAR data.
Figure 1.
The basic information of the Radarsat-2 fully polarimetric synthetic aperture radar (PolSAR) data in the Yushu study site: (a) Google Earth image, showing the coverage of the PolSAR data (red rectangle) and the location of Yushu County (yellow rectangle); (b) the Pauli RGB image of the PolSAR data.
Figure 2.
The block-level reference map of building damage in the Yushu study site.
Figure 2.
The block-level reference map of building damage in the Yushu study site.
Figure 3.
The Pauli RGB image of the ALOS-1 PolSAR data in Ishinomaki city.
Figure 3.
The Pauli RGB image of the ALOS-1 PolSAR data in Ishinomaki city.
Figure 4.
The ground-truth map interpreted by Tohoku University and the University of Tokyo [
36].
Figure 4.
The ground-truth map interpreted by Tohoku University and the University of Tokyo [
36].
Figure 5.
The block-level reference map of building damage in the Ishinomaki study site.
Figure 5.
The block-level reference map of building damage in the Ishinomaki study site.
Figure 6.
The Pauli RGB image of the ALOS-2 PolSAR data in Mashiki town.
Figure 6.
The Pauli RGB image of the ALOS-2 PolSAR data in Mashiki town.
Figure 7.
The three-grade grid-level reference map of building damage in Mashiki town.
Figure 7.
The three-grade grid-level reference map of building damage in Mashiki town.
Figure 8.
Flowchart of the proposed building damage detection method. PolSAR, polarimetric synthetic aperture radar; OPCE, optimization of polarimetric contrast enhancement; GLCM, gray level co-occurrence matrix; RVI, the radar vegetation index (RVI); SEI, the intensity component of the Shannon entropy.
Figure 8.
Flowchart of the proposed building damage detection method. PolSAR, polarimetric synthetic aperture radar; OPCE, optimization of polarimetric contrast enhancement; GLCM, gray level co-occurrence matrix; RVI, the radar vegetation index (RVI); SEI, the intensity component of the Shannon entropy.
Figure 9.
Scatter diagrams of the π/4 double-bounce scattering component of the Pauli decomposition for the non-building samples and built-up area samples in the Yushu study site.
Figure 9.
Scatter diagrams of the π/4 double-bounce scattering component of the Pauli decomposition for the non-building samples and built-up area samples in the Yushu study site.
Figure 10.
Built-up area samples (red) and non-building area samples (water samples—blue, road samples—purple, bare soil samples—brown, mountain vegetation samples—green, farmland samples—orange) in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 10.
Built-up area samples (red) and non-building area samples (water samples—blue, road samples—purple, bare soil samples—brown, mountain vegetation samples—green, farmland samples—orange) in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 11.
The probability density function (pdf) of the Pauli π/4 feature of six kinds of samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 11.
The probability density function (pdf) of the Pauli π/4 feature of six kinds of samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 12.
The pdf of the radar vegetation index (RVI) of mountain vegetation, farmland, and built-up area samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 12.
The pdf of the radar vegetation index (RVI) of mountain vegetation, farmland, and built-up area samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 13.
The pdf of the intensity component of the Shannon entropy (SEI) of farmland and built-up area samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 13.
The pdf of the intensity component of the Shannon entropy (SEI) of farmland and built-up area samples in different study sites: (a) Ishinomaki study site; (b) Mashiki town study site.
Figure 14.
Optical images (data source: Google earth) of: (a) collapsed buildings; (b) orthogonally-oriented buildings; (c) obliquely-oriented buildings.
Figure 14.
Optical images (data source: Google earth) of: (a) collapsed buildings; (b) orthogonally-oriented buildings; (c) obliquely-oriented buildings.
Figure 15.
The pdfs of the circular polarization correlation coefficient feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 15.
The pdfs of the circular polarization correlation coefficient feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 16.
The pdfs of the double-bounce scattering component of the Yamaguchi decomposition feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 16.
The pdfs of the double-bounce scattering component of the Yamaguchi decomposition feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 17.
The pdfs of the total power (Span) feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 17.
The pdfs of the total power (Span) feature: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 18.
The optical images (data source: Google earth) of the special seriously damaged areas with a few typical orthogonally-oriented buildings (red circles).
Figure 18.
The optical images (data source: Google earth) of the special seriously damaged areas with a few typical orthogonally-oriented buildings (red circles).
Figure 19.
The schematic diagram of the OPCE matching algorithm.
Figure 19.
The schematic diagram of the OPCE matching algorithm.
Figure 20.
The results of the
feature and the samples in three study sites: (
a) the samples in the Yushu study site; (
b) the result of the
feature in the Yushu study site; (
c) the samples in the Ishinomaki study site; (
d) the result of the
feature in the Ishinomaki study site; (
e) the samples in the Mashiki town study site; (
f) the result of the
feature in the Mashiki town study site. In subfigures (
a), (
c), and (
e), yellow rectangles show the target sample sets used for the OPCE matching algorithm; red patches show the collapsed building samples used to draw the pdfs of collapsed buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23; green patches show the obliquely-oriented building samples used for drawing pdfs of obliquely-oriented buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23; and blue patches show the orthogonally-oriented building samples used for drawing pdfs of orthogonally-oriented buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23.
Figure 20.
The results of the
feature and the samples in three study sites: (
a) the samples in the Yushu study site; (
b) the result of the
feature in the Yushu study site; (
c) the samples in the Ishinomaki study site; (
d) the result of the
feature in the Ishinomaki study site; (
e) the samples in the Mashiki town study site; (
f) the result of the
feature in the Mashiki town study site. In subfigures (
a), (
c), and (
e), yellow rectangles show the target sample sets used for the OPCE matching algorithm; red patches show the collapsed building samples used to draw the pdfs of collapsed buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23; green patches show the obliquely-oriented building samples used for drawing pdfs of obliquely-oriented buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23; and blue patches show the orthogonally-oriented building samples used for drawing pdfs of orthogonally-oriented buildings in
Figure 15,
Figure 16 and
Figure 17 and
Figure 21,
Figure 22 and
Figure 23.
Figure 21.
The pdfs of the feature in the Yushu study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 21.
The pdfs of the feature in the Yushu study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 22.
The pdfs of the feature in the Ishinomaki study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 22.
The pdfs of the feature in the Ishinomaki study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 23.
The pdfs of the feature in the Mashiki town study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 23.
The pdfs of the feature in the Mashiki town study site: (a) pdfs of collapsed buildings and orthogonally-oriented buildings; (b) pdfs of collapsed buildings and obliquely-oriented buildings.
Figure 24.
The classification results of collapsed and standing buildings in the Yushu study site: (a) only using the feature ; (b) using the feature and eight GLCM texture features together.
Figure 24.
The classification results of collapsed and standing buildings in the Yushu study site: (a) only using the feature ; (b) using the feature and eight GLCM texture features together.
Figure 25.
The results of non-building area removal in Yushu with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 25.
The results of non-building area removal in Yushu with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 26.
The results of non-building area removal in Ishinomaki with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 26.
The results of non-building area removal in Ishinomaki with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 27.
The results of non-building area removal in Mashiki town with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 27.
The results of non-building area removal in Mashiki town with different methods: (a) the method; (b) the classification method; (c) the proposed method.
Figure 28.
The detection results of the proposed method in different study sites: (a) the detection result of the Radarsat-2 PolSAR data in the Yushu study site; (b) the detection result of the ALOS-1 PolSAR data in the Ishinomaki study site; (c) the detection result of the ALOS-2 PolSAR data in the Mashiki town study site.
Figure 28.
The detection results of the proposed method in different study sites: (a) the detection result of the Radarsat-2 PolSAR data in the Yushu study site; (b) the detection result of the ALOS-1 PolSAR data in the Ishinomaki study site; (c) the detection result of the ALOS-2 PolSAR data in the Mashiki town study site.
Figure 29.
Block-level reference map and damage maps in Yushu: (a) reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 29.
Block-level reference map and damage maps in Yushu: (a) reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 30.
Block-level reference map and damage maps in Ishinomaki: (a) reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 30.
Block-level reference map and damage maps in Ishinomaki: (a) reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 31.
Grid-level reference map and damage maps in Mashiki town. (a) Reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 31.
Grid-level reference map and damage maps in Mashiki town. (a) Reference map; (b) damage map from the method; (c) damage map from the PWMF method; (d) damage map from the proposed method.
Figure 32.
Overall accuracy (%) of building damage detection when using different window sizes to calculate texture features: (a) the results in the Yushu study site; (b) the results in the Ishinomaki study site; (c) the results in Mashiki town study site.
Figure 32.
Overall accuracy (%) of building damage detection when using different window sizes to calculate texture features: (a) the results in the Yushu study site; (b) the results in the Ishinomaki study site; (c) the results in Mashiki town study site.
Figure 33.
Overall accuracy (%) of building damage detection when using different directions to calculate texture features: (a) the results in the Yushu study site; (b) the results in the Ishinomaki study site; (c) the results in the Mashiki town study site.
Figure 33.
Overall accuracy (%) of building damage detection when using different directions to calculate texture features: (a) the results in the Yushu study site; (b) the results in the Ishinomaki study site; (c) the results in the Mashiki town study site.
Figure 34.
Diagram of two evaluation methods: (a) block-count-based evaluation method; (b) pixel-count-based evaluation method.
Figure 34.
Diagram of two evaluation methods: (a) block-count-based evaluation method; (b) pixel-count-based evaluation method.
Table 1.
Jeffreys–Matusita (J–M) distance between obliquely-oriented buildings and collapsed buildings in different features in three study sites.
Table 1.
Jeffreys–Matusita (J–M) distance between obliquely-oriented buildings and collapsed buildings in different features in three study sites.
Study Site | J–M Distance between Obliquely-Oriented Buildings and Collapsed Buildings in |
---|
| | Span | |
---|
Yushu study site | 0.034 | 0.029 | 0.253 | 1.088 |
Ishinomaki study site | 0.266 | 0.154 | 0.156 | 0.963 |
Mashiki town study site | 0.009 | 0.103 | 0.057 | 0.736 |
Table 2.
J–M distance between affected collapsed buildings and standing buildings in different polarization features.
Table 2.
J–M distance between affected collapsed buildings and standing buildings in different polarization features.
Study Site | J–M Distance between Affected Collapsed Buildings and Standing Buildings in |
---|
| | Span | |
---|
Yushu study site | 0.199 | 0.159 | 0.260 | 0.211 |
Table 3.
J–M distance between affected collapsed buildings and standing buildings in different GLCM texture features.
Table 3.
J–M distance between affected collapsed buildings and standing buildings in different GLCM texture features.
Study Site | J–M Distance between Affected Collapsed Buildings and Standing Buildings in |
---|
Span | MaxC | Mean | Var 1 | Hom 2 | Con 3 | Dis 4 | Entr 5 | SeM 6 | Cor 7 |
---|
Yushu study site | 0.260 | 0.211 | 0.638 | 0.507 | 0.550 | 0.486 | 0.537 | 0.640 | 0.681 | 0.467 |
Table 4.
Confusion matrix of built-up and non-building area classification in the three study sites with different methods.
Table 4.
Confusion matrix of built-up and non-building area classification in the three study sites with different methods.
| | | Classification Method
| Proposed Method |
---|
| | Non-Building Area | Built-Up Area | Non-Building Area | Built-Up Area | Non-Building Area | Built-Up Area |
---|
Yushu | Ground truth | | | |
Non-building area | 91 | 209 | 184 | 116 | 264 | 36 |
Built-up area | 16 | 284 | 8 | 292 | 11 | 289 |
Prod. accu. 2 | 30.0% | 94.7% | 61.3% | 97.3% | 88.0% | 96.3% |
| OA 1: 62.5% | OA: 79.3% | OA: 92.2% |
Ishinomaki | Ground truth | | | |
Non-building area | 508 | 492 | 828 | 172 | 848 | 152 |
Built-up area | 59 | 941 | 121 | 879 | 32 | 968 |
Prod. accu. | 50.8% | 94.1% | 82.8% | 87.9% | 84.8% | 96.8% |
| OA: 72.5% | OA: 85.4% | OA: 90.8% |
Mashiki town | Ground truth | | | |
Non-building area | 52 | 14 | 461 | 30 | 499 | 1 |
Built-up area | 448 | 486 | 39 | 470 | 19 | 481 |
Prod. accu. | 10.4% | 97.2% | 92.2% | 94.0% | 99.8% | 96.2% |
| OA: 53.8% | OA: 93.1% | OA: 98.0% |
Table 5.
Error rates of different methods in three study sites.
Table 5.
Error rates of different methods in three study sites.
| | | Classification Method
| Proposed Method |
---|
| | Non-Building Area | Built-Up Area | Non-Building Area | Built-Up Area | Non-Building Area | Built-Up Area |
---|
Yushu | Ground truth | | | |
Collapsed buildings | 10 | 290 | 16 | 284 | 5 | 295 |
Error rate | 3.4% | 5.3% | 1.7% |
Ishinomaki | Ground truth | | | |
Collapsed buildings | 44 | 456 | 98 | 402 | 6 | 494 |
Error rate | 8.8% | 19.6% | 1.2% |
Mashiki town | Ground truth | | | |
Collapsed buildings | 22 | 378 | 42 | 358 | 14 | 386 |
Error rate | 5.5% | 10.5% | 3.5% |
Table 6.
The number of samples for multi-feature-based collapsed and standing buildings classification in three study sites.
Table 6.
The number of samples for multi-feature-based collapsed and standing buildings classification in three study sites.
Study Sites | Collapsed Building Samples (Pixels) | Standing Building Samples (Pixels) | Total (Pixels) |
---|
Yushu | 741 | 648 | 1389 |
Ishinomaki | 620 | 924 | 1544 |
Mashiki | 537 | 560 | 1097 |
Table 7.
Accuracy evaluation results of damage maps from different methods in the Yushu study site.
Table 7.
Accuracy evaluation results of damage maps from different methods in the Yushu study site.
Method | Detection Rate of Different Damage Level (%) | OA 1 (%) |
---|
Slight Damage | Moderate Damage | Serious Damage |
---|
method | 20.3 | 28.4 | 86.0 | 49.0 |
PWMF method | 66.3 | 14.0 | 76.8 | 57.3 |
Proposed method | 86.0 | 56.2 | 97.2 | 82.3 |
Table 8.
Accuracy evaluation results of damage maps from different methods in Ishinomaki.
Table 8.
Accuracy evaluation results of damage maps from different methods in Ishinomaki.
Method | Detection Rate of Different Damage Level (%) | OA 1 (%) |
---|
Slight Damage | Moderate Damage | Serious Damage |
---|
method | 61.4 | 0.0 | 92.2 | 63.2 |
PWMF method | 64.6 | 10.3 | 54.3 | 62.7 |
Proposed method | 100.0 | 26.0 | 86.3 | 97.4 |
Table 9.
Accuracy evaluation results of damage maps from different methods in Mashiki town.
Table 9.
Accuracy evaluation results of damage maps from different methods in Mashiki town.
Method | Detection Rate of Different Damage Level (%) | OA 1 (%) |
---|
Slight Damage | Moderate Damage | Serious Damage |
---|
method | 78.0 | 18.8 | 26.8 | 63.8 |
PWMF method | 73.9 | 15.3 | 22.9 | 59.8 |
Proposed method | 88.3 | 35.7 | 64.8 | 78.5 |
Table 10.
The comparison of two evaluation methods in three study sites.
Table 10.
The comparison of two evaluation methods in three study sites.
| | Block-Count-Based Evaluation | Pixel-Count-Based Evaluation |
---|
| | The Experimental Results | The Experimental Results |
---|
| | Slight | Moderate | Serious | Slight | Moderate | Serious |
---|
Yushu | Reference | | |
Slight | 28 | 2 | 0 | 9330 | 266 | 0 |
Moderate | 11 | 9 | 4 | 2116 | 4006 | 1002 |
Serious | 0 | 6 | 22 | 0 | 1588 | 9778 |
| OA: 72.0% | OA: 82.3% |
Ishinomaki | Reference | | |
Slight | 43 | 0 | 0 | 41,6661 | 0 | 0 |
Moderate | 4 | 1 | 0 | 6205 | 2183 | 0 |
Serious | 3 | 2 | 6 | 1265 | 4845 | 38,465 |
| OA: 84.7% | OA: 97.4% |
Mashiki town | Reference | | |
Slight | 272 | 31 | 5 | 9259 | 1064 | 168 |
Moderate | 13 | 19 | 22 | 442 | 651 | 733 |
Serious | 8 | 10 | 34 | 285 | 352 | 1173 |
| OA: 78.5% | OA: 78.5% |