Detailled semantic labeling (2D) result


Name D. Marmanis et al. (Full_CRF CNN)
Affiliation DLR-DFD, Germany
Abbreviation DLR_2
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
0.909
0.023
0.045
0.020
0.003
0.000
building
0.079
0.875
0.039
0.006
0.001
0.000
low_veg
0.050
0.023
0.784
0.143
0.000
0.000
tree
0.018
0.004
0.087
0.890
0.000
0.000
car
0.367
0.042
0.009
0.009
0.573
0.000
clutter
0.703
0.266
0.010
0.008
0.013
0.000
Precision/Correctness
0.845
0.937
0.799
0.841
0.860
-nan
Recall/Completeness
0.909
0.875
0.784
0.890
0.573
0.000
F1
0.876
0.905
0.791
0.865
0.688
-nan

Overall accuracy 0.856

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
0.936
0.017
0.036
0.010
0.001
0.000
building
0.071
0.894
0.033
0.003
0.001
0.000
low_veg
0.039
0.016
0.818
0.127
0.000
0.000
tree
0.008
0.002
0.070
0.920
0.000
0.000
car
0.281
0.048
0.006
0.004
0.661
0.000
clutter
0.690
0.288
0.006
0.003
0.014
0.000
Precision/Correctness
0.873
0.955
0.833
0.872
0.901
-nan
Recall/Completeness
0.936
0.894
0.818
0.920
0.661
0.000
F1
0.903
0.923
0.825
0.895
0.763
-nan

Overall accuracy 0.885
=============================================================

Tile top_mosaic_09cm_area2_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9442
0.0081
0.0273
0.0173
0.0032
0.0000
building
0.0901
0.8763
0.0294
0.0038
0.0004
0.0000
low_veg
0.0634
0.0203
0.7483
0.1679
0.0001
0.0000
tree
0.0263
0.0053
0.0894
0.8790
0.0001
0.0000
car
0.3951
0.0178
0.0009
0.0033
0.5830
0.0000
clutter
0.9808
0.0000
0.0000
0.0192
0.0000
0.0000
Precision/Correctness
0.779
0.959
0.837
0.823
0.908
-nan
Recall/Completeness
0.944
0.876
0.748
0.879
0.583
0.000
F1
0.854
0.916
0.790
0.850
0.710
-nan

Overall accuracy 0.842

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area2_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9713
0.0049
0.0165
0.0055
0.0019
0.0000
building
0.0835
0.8951
0.0203
0.0008
0.0004
0.0000
low_veg
0.0481
0.0128
0.7804
0.1586
0.0001
0.0000
tree
0.0133
0.0021
0.0775
0.9071
0.0000
0.0000
car
0.2885
0.0198
0.0000
0.0012
0.6904
0.0000
clutter
0.9958
0.0000
0.0000
0.0042
0.0000
0.0000
Precision/Correctness
0.806
0.977
0.869
0.851
0.933
-nan
Recall/Completeness
0.971
0.895
0.780
0.907
0.690
0.000
F1
0.881
0.934
0.822
0.878
0.794
-nan

Overall accuracy 0.869

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area4_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9288
0.0514
0.0113
0.0047
0.0038
0.0000
building
0.0875
0.8993
0.0104
0.0020
0.0008
0.0000
low_veg
0.1621
0.0402
0.4998
0.2979
0.0001
0.0000
tree
0.0464
0.0091
0.0493
0.8947
0.0005
0.0000
car
0.5292
0.0172
0.0029
0.0026
0.4481
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.804
0.967
0.659
0.731
0.804
--
Recall/Completeness
0.929
0.899
0.500
0.895
0.448
--
F1
0.862
0.932
0.568
0.804
0.576
--

Overall accuracy 0.883

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area4_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9523
0.0352
0.0087
0.0016
0.0022
0.0000
building
0.0712
0.9181
0.0092
0.0006
0.0008
0.0000
low_veg
0.1186
0.0180
0.5377
0.3258
0.0000
0.0000
tree
0.0187
0.0028
0.0438
0.9348
0.0000
0.0000
car
0.4682
0.0122
0.0008
0.0017
0.5171
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.837
0.980
0.683
0.759
0.839
--
Recall/Completeness
0.952
0.918
0.538
0.935
0.517
--
F1
0.891
0.948
0.602
0.838
0.640
--

Overall accuracy 0.909

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area6_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9081
0.0382
0.0388
0.0123
0.0026
0.0000
building
0.0784
0.8750
0.0421
0.0043
0.0002
0.0000
low_veg
0.0725
0.0172
0.7445
0.1650
0.0007
0.0000
tree
0.0117
0.0020
0.0495
0.9366
0.0001
0.0000
car
0.4291
0.0350
0.0003
0.0033
0.5323
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.840
0.958
0.676
0.924
0.914
--
Recall/Completeness
0.908
0.875
0.745
0.937
0.532
--
F1
0.873
0.915
0.708
0.930
0.673
--

Overall accuracy 0.882

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area6_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9438
0.0232
0.0284
0.0039
0.0007
0.0000
building
0.0680
0.8949
0.0360
0.0010
0.0001
0.0000
low_veg
0.0577
0.0091
0.7891
0.1435
0.0005
0.0000
tree
0.0033
0.0006
0.0361
0.9601
0.0000
0.0000
car
0.3362
0.0288
0.0000
0.0005
0.6345
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.874
0.977
0.718
0.949
0.967
--
Recall/Completeness
0.944
0.895
0.789
0.960
0.634
--
F1
0.907
0.934
0.752
0.955
0.766
--

Overall accuracy 0.913

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area8_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8843
0.0241
0.0774
0.0121
0.0020
0.0000
building
0.0556
0.9089
0.0305
0.0049
0.0001
0.0000
low_veg
0.0370
0.0117
0.8917
0.0594
0.0002
0.0000
tree
0.0210
0.0041
0.0840
0.8908
0.0000
0.0000
car
0.3998
0.0030
0.0143
0.0081
0.5748
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.897
0.951
0.772
0.927
0.933
--
Recall/Completeness
0.884
0.909
0.892
0.891
0.575
--
F1
0.891
0.929
0.828
0.909
0.711
--

Overall accuracy 0.888

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area8_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9112
0.0197
0.0644
0.0039
0.0007
0.0000
building
0.0432
0.9337
0.0211
0.0019
0.0001
0.0000
low_veg
0.0233
0.0052
0.9223
0.0490
0.0001
0.0000
tree
0.0084
0.0014
0.0619
0.9283
0.0000
0.0000
car
0.3017
0.0013
0.0075
0.0054
0.6842
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.932
0.967
0.817
0.952
0.971
--
Recall/Completeness
0.911
0.934
0.922
0.928
0.684
--
F1
0.922
0.950
0.866
0.940
0.803
--

Overall accuracy 0.920

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area10_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8579
0.0471
0.0725
0.0205
0.0020
0.0000
building
0.0442
0.9147
0.0280
0.0125
0.0005
0.0000
low_veg
0.0567
0.0280
0.7647
0.1502
0.0003
0.0000
tree
0.0122
0.0056
0.0976
0.8845
0.0000
0.0000
car
0.3278
0.0175
0.0021
0.0078
0.6447
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.927
0.901
0.714
0.830
0.833
--
Recall/Completeness
0.858
0.915
0.765
0.885
0.645
--
F1
0.891
0.908
0.738
0.857
0.727
--

Overall accuracy 0.859

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area10_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8892
0.0379
0.0623
0.0094
0.0011
0.0000
building
0.0380
0.9307
0.0222
0.0088
0.0004
0.0000
low_veg
0.0546
0.0195
0.7997
0.1258
0.0004
0.0000
tree
0.0053
0.0024
0.0738
0.9184
0.0000
0.0000
car
0.2350
0.0166
0.0004
0.0023
0.7456
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.943
0.924
0.754
0.872
0.870
--
Recall/Completeness
0.889
0.931
0.800
0.918
0.746
--
F1
0.915
0.927
0.776
0.894
0.803
--

Overall accuracy 0.889

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area12_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8821
0.0295
0.0446
0.0408
0.0030
0.0000
building
0.1450
0.8324
0.0135
0.0091
0.0000
0.0000
low_veg
0.0442
0.0290
0.7432
0.1834
0.0003
0.0000
tree
0.0085
0.0067
0.0828
0.9020
0.0000
0.0000
car
0.2316
0.0009
0.0134
0.0059
0.7482
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.737
0.913
0.845
0.846
0.880
--
Recall/Completeness
0.882
0.832
0.743
0.902
0.748
--
F1
0.803
0.871
0.791
0.873
0.809
--

Overall accuracy 0.839

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area12_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9119
0.0265
0.0369
0.0240
0.0007
0.0000
building
0.1512
0.8376
0.0083
0.0029
0.0000
0.0000
low_veg
0.0422
0.0190
0.7797
0.1588
0.0003
0.0000
tree
0.0039
0.0031
0.0656
0.9274
0.0000
0.0000
car
0.1382
0.0000
0.0135
0.0002
0.8481
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.750
0.940
0.876
0.876
0.949
--
Recall/Completeness
0.912
0.838
0.780
0.927
0.848
--
F1
0.823
0.886
0.825
0.901
0.896
--

Overall accuracy 0.865

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area14_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8367
0.0261
0.0990
0.0357
0.0025
0.0000
building
0.0247
0.9380
0.0306
0.0068
0.0000
0.0000
low_veg
0.0605
0.0453
0.7477
0.1453
0.0012
0.0000
tree
0.0214
0.0108
0.1131
0.8546
0.0001
0.0000
car
0.2138
0.0000
0.0000
0.0040
0.7822
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.875
0.886
0.742
0.861
0.699
--
Recall/Completeness
0.837
0.938
0.748
0.855
0.782
--
F1
0.855
0.911
0.745
0.858
0.738
--

Overall accuracy 0.838

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area14_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8581
0.0222
0.0935
0.0243
0.0018
0.0000
building
0.0198
0.9571
0.0209
0.0022
0.0000
0.0000
low_veg
0.0550
0.0316
0.7788
0.1332
0.0013
0.0000
tree
0.0140
0.0050
0.1002
0.8808
0.0000
0.0000
car
0.1191
0.0000
0.0000
0.0001
0.8808
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.900
0.919
0.766
0.884
0.700
--
Recall/Completeness
0.858
0.957
0.779
0.881
0.881
--
F1
0.879
0.938
0.772
0.882
0.780
--

Overall accuracy 0.863

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area16_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8369
0.0197
0.0822
0.0590
0.0022
0.0000
building
0.0732
0.8727
0.0397
0.0143
0.0001
0.0000
low_veg
0.0211
0.0187
0.8231
0.1371
0.0000
0.0000
tree
0.0223
0.0057
0.0985
0.8734
0.0000
0.0000
car
0.1896
0.0058
0.0197
0.0099
0.7750
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.823
0.936
0.814
0.860
0.828
--
Recall/Completeness
0.837
0.873
0.823
0.873
0.775
--
F1
0.830
0.903
0.818
0.866
0.801
--

Overall accuracy 0.852

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area16_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8734
0.0142
0.0742
0.0373
0.0010
0.0000
building
0.0705
0.8917
0.0323
0.0055
0.0000
0.0000
low_veg
0.0171
0.0121
0.8484
0.1224
0.0000
0.0000
tree
0.0150
0.0025
0.0851
0.8973
0.0000
0.0000
car
0.1064
0.0019
0.0096
0.0012
0.8808
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.849
0.960
0.839
0.885
0.897
--
Recall/Completeness
0.873
0.892
0.848
0.897
0.881
--
F1
0.861
0.925
0.844
0.891
0.889
--

Overall accuracy 0.878

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area20_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8703
0.0222
0.0506
0.0554
0.0014
0.0000
building
0.0194
0.9352
0.0364
0.0090
0.0001
0.0000
low_veg
0.0507
0.0240
0.7408
0.1844
0.0001
0.0000
tree
0.0075
0.0011
0.0448
0.9466
0.0000
0.0000
car
0.3108
0.0082
0.0186
0.0266
0.6357
0.0000
clutter
0.1066
0.8894
0.0018
0.0022
0.0000
0.0000
Precision/Correctness
0.896
0.820
0.805
0.889
0.886
-nan
Recall/Completeness
0.870
0.935
0.741
0.947
0.636
0.000
F1
0.883
0.874
0.772
0.917
0.740
-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_09cm_area20_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9013
0.0154
0.0480
0.0347
0.0007
0.0000
building
0.0170
0.9494
0.0306
0.0029
0.0001
0.0000
low_veg
0.0435
0.0165
0.7856
0.1543
0.0001
0.0000
tree
0.0031
0.0003
0.0333
0.9634
0.0000
0.0000
car
0.2264
0.0010
0.0075
0.0155
0.7495
0.0000
clutter
0.0948
0.9041
0.0007
0.0004
0.0000
0.0000
Precision/Correctness
0.921
0.833
0.835
0.918
0.930
-nan
Recall/Completeness
0.901
0.949
0.786
0.963
0.749
0.000
F1
0.911
0.887
0.809
0.940
0.830
-nan

Overall accuracy 0.888

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area22_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8960
0.0241
0.0329
0.0432
0.0038
0.0000
building
0.0467
0.9264
0.0156
0.0112
0.0001
0.0000
low_veg
0.0798
0.0384
0.7136
0.1676
0.0006
0.0000
tree
0.0170
0.0032
0.0522
0.9276
0.0000
0.0000
car
0.4227
0.0015
0.0121
0.0233
0.5404
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.894
0.910
0.835
0.843
0.815
--
Recall/Completeness
0.896
0.926
0.714
0.928
0.540
--
F1
0.895
0.918
0.770
0.883
0.650
--

Overall accuracy 0.870

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area22_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9297
0.0172
0.0268
0.0243
0.0020
0.0000
building
0.0407
0.9451
0.0086
0.0056
0.0000
0.0000
low_veg
0.0760
0.0277
0.7596
0.1361
0.0006
0.0000
tree
0.0065
0.0009
0.0372
0.9554
0.0000
0.0000
car
0.3474
0.0005
0.0103
0.0142
0.6276
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.918
0.939
0.876
0.885
0.869
--
Recall/Completeness
0.930
0.945
0.760
0.955
0.628
--
F1
0.924
0.942
0.814
0.919
0.729
--

Overall accuracy 0.904

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area24_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8120
0.0308
0.1114
0.0415
0.0043
0.0000
building
0.0575
0.8625
0.0667
0.0133
0.0000
0.0000
low_veg
0.0147
0.0214
0.8897
0.0738
0.0004
0.0000
tree
0.0133
0.0065
0.1483
0.8320
0.0000
0.0000
car
0.3363
0.0019
0.0054
0.0069
0.6495
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.846
0.916
0.846
0.844
0.755
--
Recall/Completeness
0.812
0.863
0.890
0.832
0.650
--
F1
0.829
0.888
0.867
0.838
0.698
--

Overall accuracy 0.858

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area24_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8453
0.0205
0.1015
0.0301
0.0027
0.0000
building
0.0552
0.8740
0.0602
0.0107
0.0000
0.0000
low_veg
0.0120
0.0170
0.9076
0.0630
0.0004
0.0000
tree
0.0076
0.0032
0.1245
0.8646
0.0000
0.0000
car
0.2522
0.0000
0.0006
0.0019
0.7453
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.869
0.939
0.869
0.869
0.782
--
Recall/Completeness
0.845
0.874
0.908
0.865
0.745
--
F1
0.857
0.905
0.888
0.867
0.763
--

Overall accuracy 0.881

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area27_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9382
0.0231
0.0236
0.0078
0.0073
0.0000
building
0.1129
0.8509
0.0302
0.0053
0.0007
0.0000
low_veg
0.0778
0.0178
0.6856
0.2185
0.0003
0.0000
tree
0.0118
0.0026
0.0388
0.9468
0.0000
0.0000
car
0.3824
0.0146
0.0055
0.0094
0.5880
0.0000
clutter
0.7904
0.0761
0.1096
0.0088
0.0150
0.0000
Precision/Correctness
0.810
0.962
0.857
0.667
0.878
-nan
Recall/Completeness
0.938
0.851
0.686
0.947
0.588
0.000
F1
0.869
0.903
0.762
0.782
0.704
-nan

Overall accuracy 0.845

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area27_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9633
0.0149
0.0161
0.0032
0.0024
0.0000
building
0.1005
0.8731
0.0238
0.0020
0.0006
0.0000
low_veg
0.0608
0.0088
0.7236
0.2067
0.0001
0.0000
tree
0.0042
0.0006
0.0243
0.9709
0.0000
0.0000
car
0.3005
0.0123
0.0032
0.0075
0.6765
0.0000
clutter
0.8279
0.0667
0.0828
0.0062
0.0163
0.0000
Precision/Correctness
0.842
0.979
0.893
0.698
0.941
-nan
Recall/Completeness
0.963
0.873
0.724
0.971
0.677
0.000
F1
0.899
0.923
0.799
0.812
0.787
-nan

Overall accuracy 0.877

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area29_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9441
0.0145
0.0246
0.0147
0.0021
0.0000
building
0.0562
0.9236
0.0121
0.0081
0.0000
0.0000
low_veg
0.0513
0.0192
0.6237
0.3057
0.0001
0.0000
tree
0.0051
0.0016
0.0229
0.9703
0.0000
0.0000
car
0.3703
0.0061
0.0000
0.0141
0.6094
0.0000
clutter
0.5831
0.3603
0.0295
0.0097
0.0174
0.0000
Precision/Correctness
0.898
0.926
0.844
0.662
0.851
-nan
Recall/Completeness
0.944
0.924
0.624
0.970
0.609
0.000
F1
0.920
0.925
0.717
0.787
0.710
-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_09cm_area29_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9652
0.0093
0.0178
0.0067
0.0010
0.0000
building
0.0492
0.9395
0.0071
0.0043
0.0000
0.0000
low_veg
0.0320
0.0085
0.6658
0.2936
0.0000
0.0000
tree
0.0017
0.0004
0.0168
0.9811
0.0000
0.0000
car
0.2869
0.0059
0.0000
0.0113
0.6959
0.0000
clutter
0.5625
0.3940
0.0202
0.0043
0.0191
0.0000
Precision/Correctness
0.919
0.943
0.884
0.701
0.882
-nan
Recall/Completeness
0.965
0.939
0.666
0.981
0.696
0.000
F1
0.942
0.941
0.759
0.818
0.778
-nan

Overall accuracy 0.889

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area31_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9721
0.0077
0.0122
0.0036
0.0044
0.0000
building
0.1441
0.8415
0.0096
0.0034
0.0014
0.0000
low_veg
0.0616
0.0037
0.8969
0.0376
0.0001
0.0000
tree
0.0399
0.0016
0.0645
0.8935
0.0004
0.0000
car
0.2683
0.1930
0.0056
0.0077
0.5254
0.0000
clutter
0.7509
0.2217
0.0035
0.0135
0.0104
0.0000
Precision/Correctness
0.810
0.928
0.899
0.921
0.882
-nan
Recall/Completeness
0.972
0.841
0.897
0.894
0.525
0.000
F1
0.884
0.883
0.898
0.907
0.658
-nan

Overall accuracy 0.867

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area31_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9825
0.0047
0.0100
0.0009
0.0019
0.0000
building
0.1315
0.8611
0.0054
0.0007
0.0012
0.0000
low_veg
0.0386
0.0018
0.9277
0.0319
0.0000
0.0000
tree
0.0138
0.0003
0.0482
0.9376
0.0001
0.0000
car
0.1938
0.2194
0.0015
0.0021
0.5831
0.0000
clutter
0.7319
0.2501
0.0020
0.0045
0.0115
0.0000
Precision/Correctness
0.840
0.940
0.928
0.949
0.919
-nan
Recall/Completeness
0.982
0.861
0.928
0.938
0.583
0.000
F1
0.905
0.899
0.928
0.943
0.714
-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_09cm_area33_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9836
0.0055
0.0076
0.0028
0.0005
0.0000
building
0.1080
0.7160
0.1690
0.0040
0.0030
0.0000
low_veg
0.0808
0.0041
0.8526
0.0620
0.0004
0.0000
tree
0.0455
0.0010
0.1474
0.8061
0.0000
0.0000
car
0.3881
0.0224
0.0070
0.0004
0.5821
0.0000
clutter
0.9010
0.0685
0.0039
0.0003
0.0263
0.0000
Precision/Correctness
0.788
0.968
0.660
0.877
0.744
-nan
Recall/Completeness
0.984
0.716
0.853
0.806
0.582
0.000
F1
0.875
0.823
0.744
0.840
0.653
-nan

Overall accuracy 0.809

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area33_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9925
0.0033
0.0034
0.0008
0.0001
0.0000
building
0.0976
0.7288
0.1690
0.0018
0.0028
0.0000
low_veg
0.0496
0.0022
0.8877
0.0603
0.0002
0.0000
tree
0.0184
0.0002
0.1335
0.8479
0.0000
0.0000
car
0.2739
0.0267
0.0034
0.0000
0.6959
0.0000
clutter
0.8980
0.0723
0.0016
0.0000
0.0281
0.0000
Precision/Correctness
0.816
0.976
0.671
0.897
0.748
-nan
Recall/Completeness
0.993
0.729
0.888
0.848
0.696
0.000
F1
0.895
0.834
0.764
0.872
0.721
-nan

Overall accuracy 0.830

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area35_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.8514
0.0301
0.0930
0.0246
0.0009
0.0000
building
0.0808
0.8867
0.0245
0.0070
0.0010
0.0000
low_veg
0.0337
0.0102
0.7322
0.2235
0.0004
0.0000
tree
0.0221
0.0036
0.1197
0.8546
0.0000
0.0000
car
0.4152
0.0002
0.0008
0.0124
0.5715
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.848
0.956
0.758
0.759
0.874
--
Recall/Completeness
0.851
0.887
0.732
0.855
0.571
--
F1
0.850
0.920
0.745
0.804
0.691
--

Overall accuracy 0.830

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area35_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9016
0.0242
0.0656
0.0083
0.0003
0.0000
building
0.0693
0.9148
0.0127
0.0022
0.0010
0.0000
low_veg
0.0238
0.0035
0.7574
0.2147
0.0005
0.0000
tree
0.0075
0.0009
0.0901
0.9016
0.0000
0.0000
car
0.3160
0.0000
0.0000
0.0058
0.6782
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.885
0.974
0.821
0.788
0.897
--
Recall/Completeness
0.902
0.915
0.757
0.902
0.678
--
F1
0.893
0.943
0.788
0.841
0.772
--

Overall accuracy 0.869

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area38_class.tif, reference set: full_reference


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9162
0.0141
0.0527
0.0146
0.0024
0.0000
building
0.0527
0.8559
0.0848
0.0054
0.0011
0.0000
low_veg
0.0410
0.0305
0.8484
0.0797
0.0004
0.0000
tree
0.0174
0.0019
0.1199
0.8607
0.0000
0.0000
car
0.3325
0.0205
0.0505
0.0124
0.5841
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.887
0.944
0.800
0.858
0.844
--
Recall/Completeness
0.916
0.856
0.848
0.861
0.584
--
F1
0.901
0.898
0.823
0.859
0.691
--

Overall accuracy 0.868

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image


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

Tile top_mosaic_09cm_area38_class.tif, reference set: no_boundary


↓ predicted || reference →
imp_surf
building
low_veg
tree
car
clutter
imp_surf
0.9460
0.0116
0.0365
0.0050
0.0009
0.0000
building
0.0453
0.8820
0.0700
0.0018
0.0010
0.0000
low_veg
0.0284
0.0293
0.8772
0.0647
0.0003
0.0000
tree
0.0057
0.0004
0.0937
0.9002
0.0000
0.0000
car
0.2553
0.0227
0.0438
0.0042
0.6739
0.0000
clutter
--
--
--
--
--
--
Precision/Correctness
0.918
0.953
0.839
0.896
0.891
--
Recall/Completeness
0.946
0.882
0.877
0.900
0.674
--
F1
0.932
0.916
0.858
0.898
0.767
--

Overall accuracy 0.900

Red/green image, indicating wrongly classified pixels


Original, resized true ortho image for better comparison


Classified image