Detailled semantic labeling (2D) result


Name D. Marmanis et al. (fCRF CNN)
Affiliation DLR-DFD, Germany
Abbreviation DLR_4
Strategy (u)nsupervised, (s)upervised, (h)ybrid s


Overall statistics, reference set: full_reference


Number of non processed tiles by this participant: 0
↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.890
0.041
0.045
0.021
0.003
0.000
building
0.037
0.930
0.025
0.007
0.000
0.000
low_veg
0.046
0.036
0.743
0.175
0.000
0.000
tree
0.016
0.006
0.078
0.900
0.000
0.000
car
0.340
0.069
0.008
0.009
0.574
0.000
clutter
0.582
0.384
0.015
0.009
0.010
0.000
Precision/Correctness
0.883
0.907
0.813
0.817
0.883
-nan
Recall/Completeness
0.890
0.930
0.743
0.900
0.574
0.000
F1
0.886
0.919
0.776
0.856
0.696
-nan

Overall accuracy 0.859

Overall statistics, reference set: no_boundary


Number of non processed tiles by this participant: 0
↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.917
0.034
0.036
0.011
0.001
0.000
building
0.029
0.949
0.018
0.004
0.000
0.000
low_veg
0.035
0.030
0.774
0.161
0.000
0.000
tree
0.007
0.003
0.063
0.928
0.000
0.000
car
0.243
0.078
0.006
0.006
0.667
0.000
clutter
0.567
0.408
0.012
0.003
0.011
0.000
Precision/Correctness
0.914
0.924
0.849
0.845
0.925
-nan
Recall/Completeness
0.917
0.949
0.774
0.928
0.667
0.000
F1
0.915
0.937
0.810
0.885
0.775
-nan

Overall accuracy 0.887
=============================================================

Tile top_mosaic_area2_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9416
0.0126
0.0281
0.0148
0.0028
0.0000
building
0.0308
0.9421
0.0236
0.0035
0.0000
0.0000
low_veg
0.0762
0.0141
0.7310
0.1786
0.0000
0.0000
tree
0.0276
0.0045
0.0744
0.8934
0.0001
0.0000
car
0.3721
0.0163
0.0004
0.0039
0.6074
0.0000
clutter
0.9828
0.0000
0.0000
0.0172
0.0000
0.0000
Precision/Correctness
0.806
0.964
0.853
0.819
0.930
-nan
Recall/Completeness
0.942
0.942
0.731
0.893
0.607
0.000
F1
0.869
0.953
0.787
0.855
0.735
-nan

Overall accuracy 0.856

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area2_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9673
0.0088
0.0175
0.0046
0.0018
0.0000
building
0.0229
0.9627
0.0139
0.0005
0.0000
0.0000
low_veg
0.0609
0.0075
0.7614
0.1702
0.0000
0.0000
tree
0.0131
0.0021
0.0610
0.9238
0.0000
0.0000
car
0.2531
0.0173
0.0000
0.0026
0.7270
0.0000
clutter
0.9970
0.0000
0.0000
0.0030
0.0000
0.0000
Precision/Correctness
0.838
0.980
0.889
0.845
0.952
-nan
Recall/Completeness
0.967
0.963
0.761
0.924
0.727
0.000
F1
0.898
0.971
0.820
0.883
0.824
-nan

Overall accuracy 0.884

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area4_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9089
0.0722
0.0105
0.0047
0.0036
0.0000
building
0.0426
0.9460
0.0087
0.0025
0.0003
0.0000
low_veg
0.1883
0.0405
0.4665
0.3046
0.0000
0.0000
tree
0.0482
0.0155
0.0410
0.8943
0.0009
0.0000
car
0.4413
0.0080
0.0002
0.0009
0.5497
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.866
0.957
0.677
0.724
0.866
--
Recall/Completeness
0.909
0.946
0.467
0.894
0.550
--
F1
0.887
0.952
0.552
0.800
0.672
--

Overall accuracy 0.904

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area4_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9331
0.0552
0.0084
0.0016
0.0016
0.0000
building
0.0274
0.9635
0.0076
0.0012
0.0002
0.0000
low_veg
0.1439
0.0157
0.5108
0.3295
0.0000
0.0000
tree
0.0199
0.0070
0.0350
0.9380
0.0001
0.0000
car
0.3515
0.0048
0.0000
0.0003
0.6434
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.909
0.971
0.707
0.754
0.923
--
Recall/Completeness
0.933
0.964
0.511
0.938
0.643
--
F1
0.921
0.967
0.593
0.836
0.758
--

Overall accuracy 0.931

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area6_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8867
0.0555
0.0419
0.0138
0.0021
0.0000
building
0.0448
0.9362
0.0135
0.0054
0.0000
0.0000
low_veg
0.0620
0.0324
0.7268
0.1786
0.0002
0.0000
tree
0.0140
0.0032
0.0405
0.9424
0.0000
0.0000
car
0.4104
0.0569
0.0031
0.0048
0.5248
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.875
0.941
0.743
0.917
0.936
--
Recall/Completeness
0.887
0.936
0.727
0.942
0.525
--
F1
0.881
0.939
0.735
0.930
0.673
--

Overall accuracy 0.896

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area6_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9211
0.0413
0.0324
0.0047
0.0005
0.0000
building
0.0339
0.9549
0.0094
0.0017
0.0000
0.0000
low_veg
0.0471
0.0221
0.7769
0.1538
0.0001
0.0000
tree
0.0049
0.0010
0.0305
0.9637
0.0000
0.0000
car
0.3119
0.0554
0.0002
0.0024
0.6301
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.914
0.961
0.788
0.945
0.984
--
Recall/Completeness
0.921
0.955
0.777
0.964
0.630
--
F1
0.918
0.958
0.782
0.954
0.768
--

Overall accuracy 0.928

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area8_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8752
0.0360
0.0747
0.0126
0.0015
0.0000
building
0.0248
0.9477
0.0223
0.0051
0.0001
0.0000
low_veg
0.0456
0.0121
0.8831
0.0590
0.0002
0.0000
tree
0.0201
0.0057
0.0650
0.9092
0.0000
0.0000
car
0.3820
0.0250
0.0160
0.0076
0.5694
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.914
0.934
0.796
0.928
0.948
--
Recall/Completeness
0.875
0.948
0.883
0.909
0.569
--
F1
0.894
0.941
0.837
0.919
0.711
--

Overall accuracy 0.896

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area8_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9015
0.0312
0.0631
0.0038
0.0003
0.0000
building
0.0137
0.9716
0.0123
0.0023
0.0001
0.0000
low_veg
0.0316
0.0058
0.9154
0.0473
0.0000
0.0000
tree
0.0076
0.0027
0.0457
0.9440
0.0000
0.0000
car
0.2743
0.0276
0.0108
0.0048
0.6826
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.950
0.951
0.840
0.954
0.985
--
Recall/Completeness
0.902
0.972
0.915
0.944
0.683
--
F1
0.925
0.961
0.876
0.949
0.806
--

Overall accuracy 0.928

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area10_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8355
0.0814
0.0572
0.0245
0.0014
0.0000
building
0.0159
0.9557
0.0155
0.0129
0.0000
0.0000
low_veg
0.0437
0.0343
0.7482
0.1738
0.0000
0.0000
tree
0.0117
0.0067
0.0651
0.9165
0.0000
0.0000
car
0.3336
0.0219
0.0034
0.0093
0.6318
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.951
0.856
0.774
0.815
0.892
--
Recall/Completeness
0.836
0.956
0.748
0.917
0.632
--
F1
0.889
0.903
0.761
0.863
0.740
--

Overall accuracy 0.864

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area10_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8675
0.0722
0.0479
0.0118
0.0006
0.0000
building
0.0116
0.9688
0.0102
0.0093
0.0000
0.0000
low_veg
0.0399
0.0254
0.7887
0.1460
0.0000
0.0000
tree
0.0049
0.0030
0.0475
0.9446
0.0000
0.0000
car
0.2383
0.0223
0.0010
0.0026
0.7358
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.966
0.877
0.814
0.857
0.947
--
Recall/Completeness
0.868
0.969
0.789
0.945
0.736
--
F1
0.914
0.921
0.801
0.899
0.828
--

Overall accuracy 0.893

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area12_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8700
0.0318
0.0527
0.0429
0.0025
0.0000
building
0.1516
0.8316
0.0101
0.0066
0.0000
0.0000
low_veg
0.0295
0.0268
0.7587
0.1849
0.0001
0.0000
tree
0.0073
0.0058
0.0885
0.8984
0.0000
0.0000
car
0.2143
0.0002
0.0261
0.0152
0.7441
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.750
0.917
0.839
0.845
0.907
--
Recall/Completeness
0.870
0.832
0.759
0.898
0.744
--
F1
0.805
0.872
0.797
0.871
0.817
--

Overall accuracy 0.840

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area12_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9006
0.0292
0.0448
0.0251
0.0004
0.0000
building
0.1601
0.8350
0.0028
0.0021
0.0000
0.0000
low_veg
0.0272
0.0198
0.7918
0.1612
0.0000
0.0000
tree
0.0029
0.0022
0.0727
0.9221
0.0000
0.0000
car
0.1092
0.0001
0.0245
0.0077
0.8585
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.761
0.939
0.868
0.873
0.980
--
Recall/Completeness
0.901
0.835
0.792
0.922
0.859
--
F1
0.825
0.884
0.828
0.897
0.915
--

Overall accuracy 0.865

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8473
0.0398
0.0681
0.0420
0.0027
0.0000
building
0.0071
0.9621
0.0246
0.0062
0.0000
0.0000
low_veg
0.0442
0.0494
0.7357
0.1706
0.0001
0.0000
tree
0.0203
0.0121
0.1031
0.8645
0.0000
0.0000
car
0.2375
0.0002
0.0000
0.0090
0.7533
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.907
0.868
0.775
0.842
0.755
--
Recall/Completeness
0.847
0.962
0.736
0.864
0.753
--
F1
0.876
0.912
0.755
0.853
0.754
--

Overall accuracy 0.845

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8685
0.0359
0.0634
0.0300
0.0023
0.0000
building
0.0032
0.9824
0.0129
0.0015
0.0000
0.0000
low_veg
0.0386
0.0377
0.7667
0.1570
0.0000
0.0000
tree
0.0135
0.0062
0.0914
0.8889
0.0000
0.0000
car
0.1277
0.0002
0.0000
0.0049
0.8672
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.931
0.896
0.799
0.867
0.760
--
Recall/Completeness
0.868
0.982
0.767
0.889
0.867
--
F1
0.899
0.937
0.783
0.878
0.810
--

Overall accuracy 0.870

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area16_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.7787
0.0479
0.1057
0.0660
0.0017
0.0000
building
0.0087
0.9570
0.0208
0.0135
0.0000
0.0000
low_veg
0.0104
0.0290
0.7574
0.2030
0.0001
0.0000
tree
0.0144
0.0045
0.0870
0.8941
0.0000
0.0000
car
0.1927
0.0022
0.0002
0.0313
0.7737
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.921
0.906
0.812
0.818
0.850
--
Recall/Completeness
0.779
0.957
0.757
0.894
0.774
--
F1
0.844
0.931
0.784
0.854
0.810
--

Overall accuracy 0.847

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area16_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8159
0.0438
0.0972
0.0426
0.0005
0.0000
building
0.0062
0.9791
0.0098
0.0050
0.0000
0.0000
low_veg
0.0075
0.0241
0.7786
0.1897
0.0001
0.0000
tree
0.0091
0.0015
0.0745
0.9149
0.0000
0.0000
car
0.0933
0.0003
0.0000
0.0176
0.8888
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.948
0.924
0.841
0.841
0.923
--
Recall/Completeness
0.816
0.979
0.779
0.915
0.889
--
F1
0.877
0.951
0.808
0.876
0.906
--

Overall accuracy 0.871

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area20_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8262
0.0493
0.0565
0.0654
0.0026
0.0000
building
0.0117
0.9496
0.0313
0.0074
0.0000
0.0000
low_veg
0.0274
0.0238
0.7396
0.2091
0.0000
0.0000
tree
0.0058
0.0013
0.0417
0.9512
0.0000
0.0000
car
0.1946
0.1310
0.0403
0.0244
0.6098
0.0000
clutter
0.0002
0.9956
0.0029
0.0012
0.0000
0.0000
Precision/Correctness
0.941
0.785
0.809
0.877
0.829
-nan
Recall/Completeness
0.826
0.950
0.740
0.951
0.610
0.000
F1
0.880
0.859
0.773
0.912
0.703
-nan

Overall accuracy 0.858

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area20_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8608
0.0403
0.0518
0.0453
0.0019
0.0000
building
0.0098
0.9628
0.0256
0.0018
0.0000
0.0000
low_veg
0.0227
0.0143
0.7861
0.1770
0.0000
0.0000
tree
0.0022
0.0003
0.0314
0.9662
0.0000
0.0000
car
0.0898
0.1562
0.0359
0.0142
0.7041
0.0000
clutter
0.0000
0.9990
0.0009
0.0002
0.0000
0.0000
Precision/Correctness
0.961
0.802
0.839
0.905
0.847
-nan
Recall/Completeness
0.861
0.963
0.786
0.966
0.704
0.000
F1
0.908
0.875
0.812
0.935
0.769
-nan

Overall accuracy 0.883

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area22_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8844
0.0311
0.0390
0.0417
0.0038
0.0000
building
0.0169
0.9623
0.0113
0.0094
0.0001
0.0000
low_veg
0.0689
0.0301
0.7380
0.1630
0.0001
0.0000
tree
0.0175
0.0037
0.0449
0.9339
0.0000
0.0000
car
0.3460
0.0328
0.0119
0.0158
0.5935
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.916
0.908
0.844
0.849
0.837
--
Recall/Completeness
0.884
0.962
0.738
0.934
0.593
--
F1
0.900
0.934
0.787
0.889
0.695
--

Overall accuracy 0.880

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area22_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9169
0.0243
0.0337
0.0231
0.0020
0.0000
building
0.0118
0.9798
0.0044
0.0040
0.0000
0.0000
low_veg
0.0648
0.0189
0.7839
0.1324
0.0000
0.0000
tree
0.0065
0.0013
0.0304
0.9618
0.0000
0.0000
car
0.2550
0.0412
0.0113
0.0086
0.6840
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.941
0.936
0.882
0.890
0.897
--
Recall/Completeness
0.917
0.980
0.784
0.962
0.684
--
F1
0.929
0.957
0.830
0.924
0.776
--

Overall accuracy 0.913

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area24_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.7753
0.0660
0.1032
0.0524
0.0032
0.0000
building
0.0505
0.8811
0.0589
0.0095
0.0000
0.0000
low_veg
0.0128
0.0968
0.7771
0.1130
0.0002
0.0000
tree
0.0082
0.0060
0.1269
0.8590
0.0000
0.0000
car
0.3280
0.0000
0.0033
0.0182
0.6506
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.862
0.765
0.846
0.796
0.816
--
Recall/Completeness
0.775
0.881
0.777
0.859
0.651
--
F1
0.816
0.819
0.810
0.826
0.724
--

Overall accuracy 0.817

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area24_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8093
0.0558
0.0933
0.0391
0.0025
0.0000
building
0.0504
0.8942
0.0492
0.0063
0.0000
0.0000
low_veg
0.0105
0.0976
0.7905
0.1012
0.0002
0.0000
tree
0.0037
0.0028
0.1055
0.8880
0.0000
0.0000
car
0.2050
0.0000
0.0001
0.0130
0.7818
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.882
0.776
0.871
0.819
0.826
--
Recall/Completeness
0.809
0.894
0.791
0.888
0.782
--
F1
0.844
0.831
0.829
0.852
0.803
--

Overall accuracy 0.837

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area27_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9350
0.0270
0.0234
0.0078
0.0067
0.0000
building
0.0582
0.9167
0.0173
0.0073
0.0005
0.0000
low_veg
0.1037
0.0252
0.6482
0.2228
0.0001
0.0000
tree
0.0118
0.0051
0.0287
0.9543
0.0001
0.0000
car
0.3800
0.0152
0.0032
0.0083
0.5934
0.0000
clutter
0.7978
0.0367
0.1126
0.0065
0.0464
0.0000
Precision/Correctness
0.842
0.956
0.883
0.662
0.887
-nan
Recall/Completeness
0.935
0.917
0.648
0.954
0.593
0.000
F1
0.886
0.936
0.748
0.781
0.711
-nan

Overall accuracy 0.861

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area27_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9597
0.0183
0.0171
0.0029
0.0020
0.0000
building
0.0455
0.9381
0.0122
0.0039
0.0004
0.0000
low_veg
0.0861
0.0155
0.6884
0.2100
0.0000
0.0000
tree
0.0047
0.0019
0.0157
0.9776
0.0000
0.0000
car
0.2912
0.0150
0.0020
0.0064
0.6854
0.0000
clutter
0.8256
0.0256
0.0992
0.0033
0.0463
0.0000
Precision/Correctness
0.879
0.973
0.918
0.693
0.950
-nan
Recall/Completeness
0.960
0.938
0.688
0.978
0.685
0.000
F1
0.917
0.955
0.787
0.811
0.796
-nan

Overall accuracy 0.893

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area29_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9335
0.0252
0.0235
0.0146
0.0032
0.0000
building
0.0166
0.9639
0.0118
0.0077
0.0000
0.0000
low_veg
0.0629
0.0242
0.5849
0.3280
0.0001
0.0000
tree
0.0038
0.0018
0.0192
0.9752
0.0000
0.0000
car
0.3508
0.0101
0.0047
0.0127
0.6217
0.0000
clutter
0.5827
0.3593
0.0283
0.0187
0.0110
0.0000
Precision/Correctness
0.916
0.912
0.843
0.650
0.831
-nan
Recall/Completeness
0.934
0.964
0.585
0.975
0.622
0.000
F1
0.925
0.937
0.691
0.780
0.711
-nan

Overall accuracy 0.863

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area29_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9544
0.0194
0.0172
0.0067
0.0022
0.0000
building
0.0103
0.9790
0.0069
0.0037
0.0000
0.0000
low_veg
0.0412
0.0125
0.6305
0.3158
0.0000
0.0000
tree
0.0010
0.0004
0.0135
0.9850
0.0000
0.0000
car
0.2625
0.0086
0.0052
0.0096
0.7142
0.0000
clutter
0.5688
0.3850
0.0248
0.0096
0.0118
0.0000
Precision/Correctness
0.939
0.930
0.882
0.688
0.851
-nan
Recall/Completeness
0.954
0.979
0.630
0.985
0.714
0.000
F1
0.947
0.954
0.735
0.810
0.776
-nan

Overall accuracy 0.892

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area31_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9366
0.0445
0.0115
0.0037
0.0036
0.0000
building
0.0340
0.9503
0.0108
0.0035
0.0014
0.0000
low_veg
0.0449
0.0064
0.9244
0.0244
0.0000
0.0000
tree
0.0371
0.0037
0.0863
0.8726
0.0003
0.0000
car
0.2152
0.2879
0.0056
0.0108
0.4806
0.0000
clutter
0.5128
0.4353
0.0303
0.0094
0.0122
0.0000
Precision/Correctness
0.889
0.858
0.882
0.936
0.885
-nan
Recall/Completeness
0.937
0.950
0.924
0.873
0.481
0.000
F1
0.912
0.902
0.902
0.903
0.623
-nan

Overall accuracy 0.883

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area31_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9494
0.0405
0.0081
0.0007
0.0013
0.0000
building
0.0219
0.9706
0.0054
0.0011
0.0011
0.0000
low_veg
0.0264
0.0037
0.9511
0.0189
0.0000
0.0000
tree
0.0131
0.0013
0.0696
0.9160
0.0000
0.0000
car
0.1387
0.3216
0.0012
0.0052
0.5332
0.0000
clutter
0.4676
0.4903
0.0254
0.0027
0.0140
0.0000
Precision/Correctness
0.925
0.874
0.916
0.965
0.924
-nan
Recall/Completeness
0.949
0.971
0.951
0.916
0.533
0.000
F1
0.937
0.920
0.933
0.940
0.676
-nan

Overall accuracy 0.911

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area33_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9383
0.0454
0.0126
0.0030
0.0007
0.0000
building
0.0438
0.8608
0.0768
0.0175
0.0011
0.0000
low_veg
0.0718
0.0230
0.4618
0.4434
0.0001
0.0000
tree
0.0380
0.0339
0.1238
0.8042
0.0000
0.0000
car
0.3395
0.1040
0.0119
0.0007
0.5439
0.0000
clutter
0.7040
0.2752
0.0029
0.0002
0.0177
0.0000
Precision/Correctness
0.837
0.859
0.638
0.531
0.816
-nan
Recall/Completeness
0.938
0.861
0.462
0.804
0.544
0.000
F1
0.885
0.860
0.536
0.639
0.653
-nan

Overall accuracy 0.773

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area33_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9515
0.0403
0.0073
0.0006
0.0003
0.0000
building
0.0360
0.8723
0.0748
0.0159
0.0011
0.0000
low_veg
0.0439
0.0169
0.4589
0.4803
0.0000
0.0000
tree
0.0150
0.0325
0.1093
0.8432
0.0000
0.0000
car
0.2311
0.1179
0.0076
0.0002
0.6432
0.0000
clutter
0.6896
0.2896
0.0016
0.0000
0.0192
0.0000
Precision/Correctness
0.868
0.876
0.654
0.526
0.820
-nan
Recall/Completeness
0.951
0.872
0.459
0.843
0.643
0.000
F1
0.908
0.874
0.539
0.648
0.721
-nan

Overall accuracy 0.791

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area35_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8451
0.0379
0.0937
0.0221
0.0012
0.0000
building
0.0359
0.9354
0.0212
0.0069
0.0006
0.0000
low_veg
0.0345
0.0088
0.7382
0.2186
0.0000
0.0000
tree
0.0250
0.0042
0.1132
0.8576
0.0001
0.0000
car
0.4242
0.0024
0.0000
0.0071
0.5663
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.886
0.952
0.767
0.765
0.897
--
Recall/Completeness
0.845
0.935
0.738
0.858
0.566
--
F1
0.865
0.943
0.752
0.809
0.694
--

Overall accuracy 0.843

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area35_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8926
0.0337
0.0665
0.0067
0.0005
0.0000
building
0.0257
0.9633
0.0086
0.0019
0.0005
0.0000
low_veg
0.0227
0.0032
0.7617
0.2123
0.0000
0.0000
tree
0.0085
0.0008
0.0845
0.9062
0.0000
0.0000
car
0.3147
0.0016
0.0000
0.0017
0.6820
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.930
0.967
0.830
0.793
0.936
--
Recall/Completeness
0.893
0.963
0.762
0.906
0.682
--
F1
0.911
0.965
0.795
0.846
0.789
--

Overall accuracy 0.882

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area38_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8930
0.0258
0.0623
0.0168
0.0021
0.0000
building
0.0215
0.9118
0.0565
0.0099
0.0004
0.0000
low_veg
0.0341
0.0537
0.7933
0.1188
0.0001
0.0000
tree
0.0156
0.0035
0.1154
0.8655
0.0000
0.0000
car
0.3299
0.0378
0.0362
0.0182
0.5779
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.919
0.909
0.805
0.807
0.890
--
Recall/Completeness
0.893
0.912
0.793
0.865
0.578
--
F1
0.906
0.911
0.799
0.835
0.701
--

Overall accuracy 0.862

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


=============================================================

Tile top_mosaic_area38_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9226
0.0235
0.0471
0.0061
0.0007
0.0000
building
0.0139
0.9385
0.0415
0.0057
0.0003
0.0000
low_veg
0.0196
0.0549
0.8186
0.1068
0.0000
0.0000
tree
0.0048
0.0012
0.0916
0.9025
0.0000
0.0000
car
0.2396
0.0419
0.0295
0.0127
0.6764
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.955
0.918
0.846
0.839
0.944
--
Recall/Completeness
0.923
0.939
0.819
0.902
0.676
--
F1
0.938
0.928
0.832
0.870
0.788
--

Overall accuracy 0.893

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image