-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
634 lines (602 loc) · 46.7 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>DiRS</title>
<!-- <link rel="stylesheet" href="http://cdn.static.runoob.com/libs/bootstrap/3.3.7/css/bootstrap.min.css"> -->
<link rel="stylesheet" href="bootstrap/css/bootstrap.min.css">
<!-- <link rel="stylesheet" type="text/css" href="css/mystyle.css"> -->
<script src="http://cdn.static.runoob.com/libs/jquery/2.1.1/jquery.min.js"></script>
<link rel="stylesheet" href="bootstrap/js/bootstrap.min.js">
</head>
<body>
<div class="container">
<div class="content">
<h1 style="text-align:center; margin-top:60px; font-weight: bold">
On Creating Benchmark Dataset for Aerial Image <br> Interpretation: Reviews, Guidances and Million-AID
</h1>
<p style="text-align:center; margin-bottom:15px; margin-top:20px; font-size: 18px">
<a href="http://www.captain-whu.com/longyang_En.html" target="_blank">Yang Long<sup>1</sup></a>,
<a href="http://www.captain-whu.com/xia_En.html" target="_blank">Gui-Song Xia<sup>1,2,*</sup></a>,
<a href="http://people.ucas.ac.cn/~shyli" target="_blank">Shengyang Li<sup>3</sup></a>,
<a href="http://www.captain-whu.com/yangwen.html" target="_blank">Wen Yang<sup>1,4</sup></a>, <br>
<a href="https://sites.google.com/site/michaelyingyang/home" target="_blank">Michael Ying Yang<sup>5</sup></a>,
<a href="https://www.sipeo.bgu.tum.de/team/zhu" target="_blank">Xiao Xiang Zhu<sup>6</sup></a>,
<a href="http://www.lmars.whu.edu.cn/prof_web/zhangliangpei/rs/index.html" target="_blank">Liangpei Zhang<sup>1</sup></a>,
<a href="http://www.lmars.whu.edu.cn/prof_web/prof_lideren/ldryinwenjl.htm" target="_blank">Deren Li<sup>1</sup></a>.
</p>
<p style="text-align:center; margin-bottom:15px; margin-top:20px; font-size: 15px;font-style: italic;">
1. State Key Lab. LIESMARS, Wuhan University, Wuhan 430079, China <br>
2. School of Computer Science, Wuhan University, Wuhan 430079, China <br>
3. Key Laboratory of Space Utilization, Technology and Engineering Center for Space Utilization, Chinese Academy of Sciences, Beijing 100094, China <br>
4. School of Electronic Information, Wuhan University, Wuhan 430072, China <br>
5. Faculty of Geo-Information Science and Earth Observation, University of Twente, Hengelosestraat 99, Enschede, Netherlands <br>
6. German Aerospace Center (DLR) and also Technical University of Munich, Germany
</p>
</div>
<!-- <br><hr> -->
<div class="row">
<div class="span6 offset2">
<ul class="nav nav-tabs">
<br />
</ul>
</div>
</div>
<table style= "width:100%;" align="center">
<tr>
<td style="text-align: center;"><a href="#DiRS" ><img src="files/DiRS-Head.jpg" class="img-responsive center-block"/> <br> DiRS</a></td>
<td style="text-align: center;"><a href="files/Million-AID.jpg" target="_blank"><img src="files/Million-AID-Head.jpg" class="img-responsive center-block" /> <br> Million-AID</a></td>
<td style="text-align: center;"><a href="files/Paper.pdf" target="_blank"><img src="files/Paper-Head.jpg" class="img-responsive center-block" /> <br> Paper</a></td>
<td style="text-align: center;"><a href="files/PPT.pdf" target="_blank" ><img src="files/PPT-Head.jpg" class="img-responsive center-block" /> <br> PPT</a></td>
</tr>
</table>
<div class="row">
<div class="span12">
<h3 style="text-align:left; margin-bottom:10px; margin-top:20px; font-weight: bold">
1. Abstract
</h3>
<p style="text-align:justify; font-size: 17px">
The past years have witnessed great progress on remote sensing (RS) image interpretation and its
wide applications. With RS images becoming more accessible than ever before, there is an increasing
demand for the automatic interpretation of these images. In this context, the benchmark datasets
serve as essential prerequisites for developing and testing intelligent interpretation algorithms.
After reviewing existing benchmark datasets in the research community of RS image interpretation,
this article discusses the problem of how to efficiently prepare a suitable benchmark dataset for
RS image interpretation. Specifically, we first analyze the current challenges of developing
intelligent algorithms for RS image interpretation with bibliometric investigations. We then present
the general guidances on creating benchmark datasets in efficient manners. Following the presented
guidances, we also provide an example on building RS image dataset, <i>i.e.</i>,
<a href="#Million-AID"><strong>Million-AID</strong></a>, a new large-scale benchmark dataset containing
a million instances for RS image scene classification. Several challenges and perspectives in RS image
annotation are finally discussed to facilitate the research in benchmark dataset construction. We do
hope this paper will provide the RS community an overall perspective on constructing large-scale and
practical image datasets for further research, especially data-driven ones.
<!-- The past decade has witnessed the great progress on remote sensing (RS) image interpretation and
its wide applications. With RS images becoming more accessible than ever before, there is an
increasing demand for the automatic interpretation of these images, where benchmark datasets are
essential prerequisites for developing and testing intelligent interpretation algorithms. After
reviewing existing benchmark datasets in the research community of RS image interpretation, this
article discusses the problem of how to efficiently prepare a suitable benchmark dataset for RS
image analysis. Specifically, we first analyze the current challenges of developing intelligent
algorithms for RS image interpretation with bibliometric investigations. We then present some
principles, <i>i.e.</i>, <i><strong>di</strong>versity</i>, <i><strong>r</strong>ichness</i>,
and <i><strong>s</strong>calability</i> (called <a href="#DiRS"><strong>DiRS</strong></a>), on
constructing benchmark datasets in efficient manners. Following the DiRS principles, we also
provide an example on building datasets for RS image classification, <i>i.e.</i>,
<a href="#Million-AID"><strong>Million-AID</strong></a>, a new large-scale benchmark dataset
containing million instances for RS scene classification. Several challenges and perspectives in
RS image annotation are finally discussed to facilitate the research in benchmark dataset construction.
We do hope this paper will provide RS community an overall perspective on constructing large-scale
and practical image datasets for further research, especially data-driven ones. -->
</p>
</div>
</div>
<br>
<div class="row">
<div class="span12">
<h3 style="text-align:left; margin-bottom:10px; margin-top:20px; font-weight: bold">
2. Annotated Datasets for RS Image Interpretation
</h3>
<p style="text-align:justify; font-size: 17px">
The interpretation of RS images has been playing an increasingly important role in a large
diversity of applications, and thus, has attracted remarkable research attentions. Consequently,
various datasets have been built to advance the development of interpretation algorithms for RS
images. Covering literature published over the past decade, we perform a systematic review of
the existing RS image datasets concerning the current mainstream of RS image interpretation tasks,
including scene classification, object detection, semantic segmentation and change detection.
</p>
<!-- <h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Tag Cloud of RS Image Interpretation
</h4>
<img src="files/TagCloud.jpg" width="500" class="img-responsive center-block" /> -->
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Scene Classification
</h4>
<table class="table .table-condensed"
style="border-top: solid 1px rgb(90, 90, 90); border-bottom: solid 1px rgb(90, 90, 90); font-family: Times; font-size: small; width: 75%;"
align="center" frame="hsides">
<caption style="color:black; text-align: center;font-size: 17px;">
Comparison among different RS image scene classification datasets
</caption>
<thead>
<tr style="border-bottom: solid 1px rgb(120, 120, 120);">
<th style="text-align: left;">Dataset</th>
<th style="text-align: center;">#Cat.</th>
<th style="text-align: center;">#Images per cat.</th>
<th style="text-align: center;">#Images</th>
<th style="text-align: center;">Resolution (m)</th>
<th style="text-align: center;">Image size</th>
<th style="text-align: center;">GL/IT/SP</th>
<th style="text-align: center;">Year</th>
</tr>
</thead>
<tbody>
<tr style="border-top: solid 1px white">
<td style="text-align: left;">
<a href="http://weegee.vision.ucmerced.edu/datasets/landuse.html" target="_blank">UC-Merced</a> <br>
<a href="https://captain-whu.github.io/BED4RS/" target="_blank"> WHU-RS19 </a> <br>
<a href="https://docs.google.com/viewer?a=v&pid=sites&srcid=ZGVmYXVsdGRvbWFpbnxxaW56b3VjbnxneDo1MDYzYWMxOWIwMjRiMWFi" target="_blank"> RSSCN7 </a> <br>
<a href="http://csc.lsu.edu/~saikat/deepsat/" target="_blank"> SAT-4 </a> <br>
<a href="http://csc.lsu.edu/~saikat/deepsat/" target="_blank"> SAT-6 </a> <br>
<a href="http://patreo.dcc.ufmg.br/2017/11/12/brazilian-coffee-scenes-dataset" target="_blank"> BCS </a> <br>
<a href="https://pan.baidu.com/s/1mhagndY" target="_blank"> RSC11 </a> <br>
<a href="http://www.lmars.whu.edu.cn/prof_web/zhongyanfei/Num/Google.html" target="_blank">SIRI-WHU <br>
<a href="http://www.escience.cn/people/JunweiHan/NWPU-RESISC45.html" target="_blank"> NWPU-RESISC45 </a> <br>
<a href="http://www.captain-whu.com/project/AID" target="_blank"> AID </a> <br>
<a href="https://github.com/lehaifeng/RSI-CB" target="_blank"> RSI-CB128 </a> <br>
<a href="https://github.com/lehaifeng/RSI-CB" target="_blank"> RSI-CB256 </a> <br>
<a href="https://www.kaggle.com/c/planet-understanding-the-amazon-from-space/overview" target="_blank"> Planet-UAS </a> <br>
<a href="https://github.com/RSIA-LIESMARS-WHU/RSD46-WHU" target="_blank"> RSD46-WHU </a> <br>
<a href="https://www.iuii.ua.es/datasets/masati/" target="_blank"> MASATI </a> <br>
<a href="https://github.com/phelber/eurosat" target="_blank"> EuroSAT</a> <br>
<a href="https://sites.google.com/view/zhouwx/dataset" target="_blank"> PatternNet </a> <br>
<a href="https://github.com/fMoW/dataset" target="_blank"> fMoW </a> <br>
<a href="https://www.kaggle.com/c/widsdatathon2019/data" target="_blank"> WiDS Datathon 2019</a><br>
<a href="https://drive.google.com/file/d/1Fk9a0DW8UyyQsR8dP2Qdakmr69NVBhq9/view" target="_blank"> Optimal-31 </a> <br>
<a href="http://bigearth.net/" target="_blank"> BigEarthNet </a> <br>
<a href="https://github.com/lehaifeng/CLRS" target="_blank"> CLRS </a> <br>
<a href="https://data.mendeley.com/datasets/7j9bv9vwsx/2" target="_blank"> MLRSN </a>
</td>
<td style="text-align: center;">21 <br> 19 <br> 7 <br> 4 <br> 6 <br> 2 <br> 11 <br> 12 <br>
45 <br> 30 <br> 45 <br> 35 <br> 17 <br> 46 <br> 7 <br> 10 <br> 38 <br> 62 <br> 2 <br> 31 <br> 43 <br>
25 <br> 46</td>
<td style="text-align: center;">100 <br> 50 to 61 <br> 400 <br> 89,963 to 178,034 <br>
10,262 to 150,400 <br> 1,438 <br> ~100 <br> 200 <br> 700 <br> 220 to 420 <br> 173 to
1,550 <br> 198 to 1,331 <br> -- <br> 500 to 3,000 <br> 304 to 1,789 <br> 2,000 to 3,000 <br> 800
<br> -- <br> -- <br> 60 <br> 328 to 217,119 <br> 600 <br> 1,500 to 3,000</td>
<td style="text-align: center;">2,100 <br> 1,013 <br> 2,800 <br> 500,000 <br> 405,000 <br>
2,876 <br> 1,232 <br> 2,400 <br> 31,500 <br> 10,000 <br> 36,000 <br> 24,000 <br> 40,408 <br> 117,000
<br> 7,389 <br> 27,000 <br> 30,400 <br> 132,716 <br> 20,000 <br> 1,860 <br> 590,326 <br> 15,000 <br> 109,161 </td>
<td style="text-align: center;">0.3 <br> up to 0.5 <br> -- <br> 1 to 6 <br> 1 to 6 <br> --
<br> ~0.2 <br> 2 <br> 0.2 to 30 <br> 0.5 to 8 <br> 0.3 to 3 <br> 0.3 to 3 <br> 3 to 5 <br> 0.5 to 2
<br> -- <br> 10 <br> 0.06 to 4.7 <br> 0.5 <br> 3 <br> -- <br> 10,20,60 <br> 0.26 to 8.85 <br> 0.1 to 10 </td>
<td style="text-align: center;">256×256 <br> 600×600 <br> 400×400 <br>
28×28 <br> 28×28 <br> 600×600 <br> 512×512 <br> 200×200
<br> 256×256 <br> 600×600 <br> 128×128 <br> 256×256 <br> 256×256 <br>
256×256 <br> 512×512 <br> 64×64 <br> 256×256 <br> 74×58 to 16184×16288
<br> 256×256<br> 256×256 <br> 20×20,60×60,120×120 <br> 256×256 <br> 256×256</td>
<td style="text-align: center;">✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br>
✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br> ✓✓✓ <br> ✗✗✗ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ </td>
<td style="text-align: center;">2010 <br> 2012 <br> 2015 <br> 2015 <br> 2015 <br> 2015 <br>
2016 <br> 2016 <br> 2016 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2018 <br> 2018 <br> 2018 <br> 2018
<br> 2019 <br> 2019 <br> 2019 <br> 2020 <br> 2020
</td>
</tr>
</tbody>
</table>
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Object Detection
</h4>
<table class="table .table-condensed"
style="border-top: solid 1px rgb(90, 90, 90); border-bottom: solid 1px rgb(90, 90, 90); font-family: Times; font-size: small; width: 75%;"
align="center" frame="hsides">
<caption style="color:black; text-align: center;font-size: 17px;">
Comparison among different RS image object detection datasets
</caption>
<thead>
<tr style="border-bottom: solid 1px rgb(120, 120, 120);">
<th style="text-align: left;">Dataset</th>
<th style="text-align: center;">#Annot.</th>
<th style="text-align: center;">#Cat.</th>
<th style="text-align: center;">#Instances</th>
<th style="text-align: center;">#Images</th>
<th style="text-align: center;">Resolution (m) </th>
<th style="text-align: center;">Image width</th>
<th style="text-align: center;">GL/IT/SP</th>
<th style="text-align: center;">Year</th>
</tr>
</thead>
<tbody>
<tr style="border-top: solid 1px white">
<td style="text-align: left;">
<a href="https://ai.stanford.edu/users/koller/Papers/Heitz+Koller:ECCV08.pdf" target="_blank"> TAS </a> <br>
<!-- <a> TAS </a> <br> -->
<a href="https://sourceforge.net/projects/oirds/" target="_blank">ORIDS</a> <br>
<a href="http://web.eee.sztaki.hu/remotesensing/building_benchmark.html" target="_blank"> SZTAKI-INRIA </a> <br>
<a href="https://gcheng-nwpu.github.io/datasets" target="_blank"> NWPU-VHR10 </a> <br>
<a href="https://www.dlr.de/eoc/en/desktopdefault.aspx/tabid-12760/22294_read-52777" target="_blank"> DLR-MVDA </a> <br>
<!-- <a href="https://ucassdl.cn/downloads/publication/ICIP2015_ZhuHaigang.pdf" target="_blank"> UCAS-AOD </a> <br> -->
<a href="https://github.com/ming71/UCAS-AOD-benchmark" target="_blank"> UCAS-AOD </a> <br>
<a href="https://downloads.greyc.fr/vedai" target="_blank"> VEDAI </a> <br>
<a href="https://gdo152.llnl.gov/cowc" target="_blank">COWC</a> <br>
<a href="https://sites.google.com/site/hrsc2016/" target="_blank"> HRSC2016 </a> <br>
<a href="https://github.com/RSIA-LIESMARS-WHU/RSOD-Dataset-" target="_blank"> RSOD </a> <br>
<a href="https://lafi.github.io/LPN/" target="_blank"> CARPK </a> <br>
<!-- <a href="https://zhuanlan.zhihu.com/p/143794468" target="_blank"> SSDD/SSDD+ </a> <br> -->
<a> SSDD/SSDD+ </a> <br>
<a href="https://spacenet.ai/datasets/" target="_blank"> SpaceNet1-6 </a> <br>
<a href="https://pan.baidu.com/s/1geTwAVD" target="_blank">LEVIR</a> <br>
<a href="https://github.com/VisDrone/VisDrone-Dataset" target="_blank">VisDrone</a> <br>
<a href="http://xviewdataset.org/#dataset" target="_blank">xView</a> <br>
<a href="https://captain-whu.github.io/DOTA/" target="_blank">DOTA-v1.0</a> <br>
<a href="http://www.graphnetcloud.cn/2-8" target="_blank">ITCVD</a> <br>
<a href="https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html" target="_blank">WHU building dataset</a> <br>
<a href="https://competitions.codalab.org/competitions/18544" target="_blank">DeepGlobe Building</a> <br>
<a href="http://www.graphnetcloud.cn/2-7" target="_blank">OpenSARShip</a> <br>
<a href="https://www.crowdai.org/challenges/mapping-challenge" target="_blank">CrowdAI Mapping Challenge</a> <br>
<a href="https://www.kaggle.com/c/airbus-ship-detection/overview" target="_blank">Airbus Ship Detection Challenge</a> <br>
<a href="https://captain-whu.github.io/iSAID/" target="_blank">iSAID</a> <br>
<a href="https://gitee.com/cofferlait/TGRS-HRRSD-Dataset" target="_blank">HRRSD</a> <br>
<a href="https://gcheng-nwpu.github.io/datasets" target="_blank">DIOR</a> <br>
<a href="https://captain-whu.github.io/DOAI2019/dataset.html" target="_blank">DOTA-v1.5</a> <br>
<a href="https://github.com/CAESAR-Radi/SAR-Ship-Dataset" target="_blank">SAR-Ship-Dataset</a> <br>
<a href="http://radars.ie.ac.cn/web/data/getData?dataType=SARDataset" target="_blank"> AIR-SARShip </a><br>
<a href="https://github.com/chaozhong2010/HRSID" target="_blank">HRSID</a> <br>
<a href="aireverie.com/rareplanes" target="_blank">RarePlanes</a> <br>
<a href="https://captain-whu.github.io/DOTA/" target="_blank">DOTA-v2.0</a>
</td>
<td style="text-align: center;">HBB <br> OBB <br> OBB <br> HBB <br> OBB <br> OBB <br> OBB <br> CP
<br> OBB <br> HBB <br> HBB <br> HBB/OBB <br> Polygon <br> HBB <br> HBB <br> HBB <br> OBB <br> HBB
<br> Polygon <br> Polygon <br> Chip <br> Polygon <br> Polygon <br> Polygon <br> HBB <br> HBB <br> OBB <br> HBB<br> HBB <br> HBB <br> Polygon<br> OBB</td>
<td style="text-align: center;">1 <br> 5 <br> 1 <br> 10 <br> 2 <br> 2 <br> 9 <br> 1 <br> 26 <br> 4
<br> 1 <br> 1 <br> 1 <br> 3 <br> 10 <br> 60 <br> 15 <br> 1 <br> 1 <br> 2 <br> 1<br> 1 <br> 1 <br> 15 <br> 13 <br> 20 <br> 16 <br> 1
<br> 1 <br> 1 <br> 1 <br> 18</td>
<td style="text-align: center;">1,319 <br> 1,800 <br> 665 <br> 3,651 <br> 14,235 <br> 14,596 <br> 3,640
<br> 32,716 <br> 2,976 <br> 6,950 <br> 89,777 <br> 2,456 <br> 859,982 <br> 11,000 <br> 54,200 <br>
1,000,000 <br> 188,282 <br> 29,088<br> 221,107 <br> 302,701 <br> 1,1346 <br> 2,910,917 <br> ~131,000 <br> 655,451 <br> 55,740 <br> 192,472 <br>
402,089 <br> 5,9535 <br> 2,040 <br> 16,951 <br> 644,258 <br> 1,793,658 </td>
<td style="text-align: center;">30 <br> 900 <br> 9 <br> 800 <br> 20 <br> 1,510 <br> 1,210 <br> 53
<br> 1,061 <br> 976 <br> 1,448 <br> 1,160 <br> -- <br> 22,000 <br> 10,209 <br> 1,413 <br> 2,806
<br> 173 <br> 25,420 <br> 24586 <br> 41 <br> 341,058 <br> 208,162 <br> 2,806 <br> 21,761 <br> 23,463 <br> 2,806 <br> 43,819 <br> 300 <br> 5,604 <br> 50,253 <br>
11,268 </td>
<td style="text-align: center;"> -- <br> up to 0.08 <br> -- <br> 0.08 to 2 <br> 0.13 <br> -- <br> 0.125 <br>
0.15 <br> -- <br> 0.3 to 3 <br> -- <br> 1 to 15 <br> up to 0.3 <br> 0.2 to 1 <br> -- <br> 0.3 <br> up to 0.3 <br>
0.1 <br> 0.075 to 2.7 <br> 0.3 <br> ~10 <br> -- <br> -- <br> up to 0.3 <br> 0.15 to 1.2 <br> 0.5 to 30 <br>
up to 0.3 <br> up to 0.3 <br> 1;3 <br> 0.5;1;3 <br> 0.3 <br> up to 0.3</td>
<td style="text-align: center;">792 <br> 256 to 640 <br> ~800 <br> ~1,000 <br> 5,616 <br> ~1,000 <br>
512/1,024 <br> 2,000 to 19,000 <br> ~1,100 <br> ~1,000 <br> 1,280 <br> ~500 <br> -- <br> 800 <br> 2,000
<br> ~3,000 <br> 800 to 4,000 <br> 3,744×5,616 <br> 512 <br> 650 <br> -- <br> 300 <br> 768 <br> 800 to 4,000 <br> 152 to 10,569 <br>
800 <br> 800 to 13,000 <br> 256 <br> 1,000 <br> 800 <br> -- <br> 800--20,000 </td>
<td style="text-align: center;">✗✗✗ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br>✗✗✓ <br> ✗✗✗ <br> ✓✗✗ <br>
✓✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✓ <br> ✗✗✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br> ✓✗✓ <br> ✗✗✗ <br>
✗✗✗ <br> ✗✗✗ <br> ✗✗✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✗ <br> ✗✗✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✓✓✓ <br> ✗✗✗</td>
<td style="text-align: center;">2008 <br> 2009 <br> 2012 <br> 2014 <br> 2015 <br> 2015 <br> 2016 <br>
2016 <br> 2016 <br> 2017 <br> 2017 <br> 2017 <br> 2018 <br> 2018 <br> 2018 <br> 2018 <br> 2018 <br>
2018 <br> 2018 <br> 2018 <br> 2018 <br> 2018 <br> 2018 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2020 <br> 2020 <br> 2020 <br> 2020</td>
</tr>
</tbody>
</table>
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Semantic Segmentation
</h4>
<table class="table .table-condensed"
style="border-top: solid 1px rgb(90, 90, 90); border-bottom: solid 1px rgb(90, 90, 90); font-family: Times; font-size: small; width: 75%;"
align="center" frame="hsides">
<caption style="color:black; text-align: center;font-size: 17px;">
Comparison among different RS image semantic segmentation datasets
</caption>
<thead>
<tr style="border-bottom: solid 1px rgb(120, 120, 120);">
<th style="text-align: left;">Dataset</th>
<th style="text-align: center;">#Cat.</th>
<th style="text-align: center;">#Images.</th>
<th style="text-align: center;">Resolution (m)</th>
<th style="text-align: center;">#Bands</th>
<th style="text-align: center;">Image size</th>
<th style="text-align: center;">GL/IT/SP</th>
<th style="text-align: center;">Year</th>
</tr>
</thead>
<tbody>
<tr style="border-top: solid 1px white">
<td style="text-align: left;">
<a href="http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes#Kennedy_Space_Center_.28KSC.29" target="_blank">Kenney Space Center</a><br>
<a href="http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes#Botswana" target="_blank">Botswana</a> <br>
<a href="http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes#Salinas" target="_blank">Salinas </a><br>
<a href="http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes#Pavia_University_scene" target="_blank">University of Pavia </a><br>
<a href="http://www.ehu.eus/ccwintco/index.php/Hyperspectral_Remote_Sensing_Scenes#Pavia_Centre_scene" target="_blank">Pavia Centre </a><br>
<a href="http://www2.isprs.org/commissions/comm3/wg4/2d-sem-label-vaihingen.html" target="_blank">ISPRS Vaihingen </a><br>
<a href="http://www2.isprs.org/commissions/comm3/wg4/2d-sem-label-potsdam.html" target="_blank">ISPRS Potsdam </a><br>
<a href="https://www.cs.toronto.edu/~vmnih/data/" target="_blank">Massachusetts Buildings </a><br>
<a href="https://www.cs.toronto.edu/~vmnih/data/" target="_blank">Massachusetts Roads </a><br>
<a href="https://purr.purdue.edu/publications/1947/1" target="_blank">Indian Pines </a><br>
<a href="https://sites.google.com/site/michelevolpiresearch/data/zurich-dataset" target="_blank">Zurich Summer </a><br>
<a href="https://www.usgs.gov/core-science-systems/nli/landsat/spatial-procedures-automated-removal-cloud-and-shadow-sparcs" target="_blank"> SPARCS Validation </a><br>
<a href="https://landsat.usgs.gov/landsat-8-cloud-cover-assessment-validation-data" target="_blank"> Biome </a><br>
<a href="https://project.inria.fr/aerialimagelabeling/" target="_blank">Inria Dataset </a><br>
<a href="http://earthvisionlab.whu.edu.cn/zm/SemanticSegmentation/index.html" target="_blank">EvLab-SS </a><br>
<a href="https://github.com/rmkemker/RIT-18" target="_blank">RIT-18 </a><br>
<a href="https://zenodo.org/record/1154821#.YAJWkdAzZPY" target="_blank"> CITY-OSM </a><br>
<a href="https://www.kaggle.com/c/dstl-satellite-imagery-feature-detection/overview" target="_blank"> Dstl-SIFD </a><br>
<a href="http://www.classic.grss-ieee.org/2017-ieee-grss-data-fusion-contest/" target="_blank"> IEEE GRSS Data Fusion Contest 2017</a><br>
<a href="http://www.classic.grss-ieee.org/2017-ieee-grss-data-fusion-contest/" target="_blank"> IEEE GRSS Data Fusion Contest 2018 </a><br>
<a href="https://github.com/ishann/aeroscapes" target="_blank"> Aeroscapes </a><br>
<a href="https://sites.google.com/view/zhouwx/dataset#h.p_hQS2jYeaFpV0" target="_blank"> DLRSD </a><br>
<a href="https://competitions.codalab.org/competitions/18468" target="_blank"> DeepGlobe Land Cover </a><br>
<!-- <a href="https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html" target="_blank">WHU Building-Aerial Imagery </a><br>
<a href="https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html" target="_blank">WHU Building-Satellite Imagery I </a><br>
<a href="https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html" target="_blank">WHU Building-Satellite Imagery II </a><br> -->
<a href="https://mediatum.ub.tum.de/1454690" target="_blank">So2Sat LC242 </a><br>
<a href="https://mediatum.ub.tum.de/1474000" target="_blank">SEN12MS </a><br>
<a href="https://github.com/SorourMo/95-Cloud-An-Extension-to-38-Cloud-Dataset" target="_blank"> 95-Cloud </a><br>
<a href="http://im.itu.edu.pk/deepcount/" target="_blank"> Shakeel et al. </a><br>
<a href="https://zenodo.org/record/1460961#.YAJco-gzaUk" target="_blank"> ALCD Cloud Masks</a><br>
<a href="https://www.dlr.de/eoc/en/desktopdefault.aspx/tabid-12760/22294_read-58694/" target="_blank"> SkyScapes </a><br>
<a href="https://github.com/dronedeploy/dd-ml-segmentation-benchmark" target="_blank"> DroneDeploy </a><br>
<a href="http://eo-learn.sentinel-hub.com/?prefix=" target="_blank"> Slovenia LULC </a><br>
<a href="http://registry.mlhub.earth/10.34911/rdnt.d2ce8i/" target="_blank"> LandConverNet </a><br>
<a href="https://www.uavid.nl" target="_blank">UAVid </a><br>
<a href="https://x-ytong.github.io/project/GID.html" target="_blank">GID </a><br>
<a href="http://landcover.ai/" target="_blank"> LandCover.ai </a><br>
<a href="https://www.agriculture-vision.com/dataset" target="_blank"> Agriculture-Vision </a><br>
<a href="https://zenodo.org/record/4172871#.YAPWU9AzZPY" target="_blank"> Sentinel-2 Cloud Mask Catalogue </a><br>
</td>
<td style="text-align: center;">13 <br> 14 <br> 16 <br> 9 <br> 9 <br> 6 <br> 6 <br> 2 <br> 2
<br> 16 <br> 8 <br> 7 <br> 4 <br> 2 <br> 10 <br> 18 <br> 3 <br> 10 <br> 17 <br> 20 <br> 11 <br> 17 <br>
<!-- 2 <br> 2 <br> 2 <br> -->
7 <br> 17 <br> 33 <br>1 <br> 1<br> 8<br>31 <br> 7 <br> 10 <br>7 <br>
8 <br> 15 <br> 3 <br> 9 <br> 18 </td>
<td style="text-align: center;">1 <br> 1 <br> 1 <br> 1 <br> 1 <br> 33 <br> 38 <br> 151 <br>
1,171 <br> 1 <br> 20 <br> 80 <br> 96 <br> 360 <br> 60 <br> 3 <br> 1,671 <br> 57 <br> 30 <br> 1 <br> 3,269 <br> 2,100 <br>
<!-- <br> 8,189 <br> 204 <br> 17,388 -->
1,146 <br> 400,673 <br>180,662 triplets <br> 43,902 <br> 2,682 <br> 38 <br> 16 <br> 55 <br> 940 <br> 1,980 <br>
420 <br>150 <br> 41 <br> 94,986 <br> 513 </td>
<td style="text-align: center;">18 <br> 30 <br> 3.7 <br> 1.3 <br> 1.3 <br> 0.09 <br> 0.05
<br> 1 <br> 1 <br> 20 <br> 0.62 <br> 30 <br> 30 <br> 0.3 <br> 0.1 to 2 <br> 0.047 <br> 0.1 <br> up to 0.3 <br> 1.4 <br> 1 <br> -- <br> 0.3 <br>
<!-- <br> 0.3 <br> 0.3 to 2.5 <br> 2.7 -->
0.5 <br> 10 <br> 10 to 50 <br> 30 <br> 0.3 <br> 10 <br> 0.13 <br> 0.1 <br> 10 <br> 10 <br>
-- <br> 0.8 to 10 <br> 0.25,0.5<br> 0.1,15,0.2 <br> 20</td>
<td style="text-align: center;">224 bands <br> 242 bands <br> 224 bands <br> 115 bands <br>
115 bands <br> IR, R, G, DSM, nDSM <br> IR, RGB, DSM, nDSM <br> RGB <br> RGB <br> 224
bands <br> NIR, RGB <br> 11 <br> 11 <br> RGB <br> RGB <br> 6 bands <br> RGB <br> up to 16 <br> 9 <br> 48 <br> RGB <br> RGB
<!-- <br> RGB <br> RGB <br> RGB -->
<br> RGB <br> 10 bands <br> up to 13 bands <br> NIR,RGB <br> RGB <br> RGB <br> RGB <br> RGB <br> 6 <br> NIR,RGB <br>
RGB <br> 4 bands <br> RGB <br> NIR,RGB <br> 13 </td>
<td style="text-align: center;">512×614 <br> 1,476×256 <br> 512×217 <br>
610×340 <br> 1,096×492 <br> ~2,500×2500 <br> 6,000×6,000 <br>
1,500×1,500 <br> 1,500×1,500 <br> 145×145 <br> 1,000×1,150 <br>
1000×1000<br> ~9000×9000 <br> 5000×5000 <br> 4,500×4,500 <br>
9,000×6,000 <br> 2500×2500 to 3300×33300 <br> ~3,350×3,400 <br> 643×666,374×515 <br> 4172×1202 <br> 720×1280 <br> 256×256
<!-- <br> 512×512 <br> 512×512 <br> 512×512 -->
<br>2,448×2,448<br> 32×32 <br> 256×256 <br> 384×384 <br> 300×300 <br> 1,830×1,830 <br> 5,616×3,744 <br> up to 12,039×13,854 <br> 5,00×5,00 <br> 256×256 <br>
~4,000×2,160 <br>6,800×7,200 <br>9,000×9,500;4,200×4,700 <br> 512×512 <br> 1,024×1,024</td>
<td style="text-align: center;">✗✓✓ <br> ✗✓✓ <br> ✗✗✓ <br> ✗✗✓ <br> ✗✗✓ <br> ✗✗✓ <br> ✓✗✓
<br> ✓✓✗ <br> ✓✓✗ <br> ✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✓✗✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br>
<!-- <br> 2017 <br> 2017 <br> 2019 <br> -->
✗✗✓ <br>✓✗✓ <br> ✓✗✓ <br> ✓✗✓ <br> ✗✗✗ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✗✓ <br>
✓✓✓ <br> ✓✗✗ <br> ✗✗✓ <br> ✓✓✓ </td>
<td style="text-align: center;">2005 <br> 2005 <br> -- <br> -- <br> -- <br> 2012 <br> 2012
<br> 2013 <br> 2013 <br> 2015 <br> 2015 <br> 2016 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2017 <br> 2018 <br> 2018 <br> 2018 <br>
<!-- <br> 2017 <br> 2017 <br> 2019 <br> -->
2018 <br>2019 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2020 <br> 2020 <br>
2020 <br> 2020 <br> 2020 <br> 2020 </td>
</tr>
</tbody>
</table>
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Change Detection
</h4>
<table class="table .table-condensed"
style="border-top: solid 1px rgb(90, 90, 90); border-bottom: solid 1px rgb(90, 90, 90); font-family: Times; font-size: small; width: 75%;"
align="center" frame="hsides">
<caption style="color:black; text-align: center;font-size: 17px;">
Comparison among different RS image change detection datasets
</caption>
<thead>
<tr style="border-bottom: solid 1px rgb(120, 120, 120);">
<th style="text-align: left;">Dataset</th>
<th style="text-align: center;">#Cat.</th>
<th style="text-align: center;">#Image pairs.</th>
<th style="text-align: center;">Resolution (m)</th>
<th style="text-align: center;">#Bands</th>
<th style="text-align: center;">Image size</th>
<th style="text-align: center;">GL/IT/SP</th>
<th style="text-align: center;">Year</th>
</tr>
</thead>
<tbody>
<tr style="border-top: solid 1px white">
<td style="text-align: left;">
<a href="http://web.eee.sztaki.hu/remotesensing/airchange_benchmark.html" target="_blank">SZTAKI AirChange </a><br>
<a href="https://computervisiononline.com/dataset/1105138664" target="_blank">AICD </a><br>
<!-- <a href="https://ieeexplore.ieee.org/abstract/document/6553145" target="_blank">Taizhou Data </a><br> -->
<a>Taizhou Data </a><br>
<!-- <a href="https://ieeexplore.ieee.org/abstract/document/6553145" target="_blank">Kunshan Data </a><br> -->
<a>Kunshan Data </a><br>
<!-- <a href="https://www.mdpi.com/2072-4292/9/3/284" target="_blank">Yangcheng</a> <br> -->
<a href="https://sites.google.com/site/michelevolpiresearch/codes/cross-sensor" target="_blank">Cross-sensor Bastrop</a><br>
<a href="http://sigma.whu.edu.cn/newspage.php?q=2019_03_26_ENG" target="_blank"> MtS-WH </a> <br>
<a>Yancheng</a><br>
<a href="https://drive.google.com/file/d/1cWy6KqE0rymSk5-ytqr7wM1yLMKLukfP/view" target="_blank">GETNET dataset</a><br>
<!-- <a href="https://ieeexplore.ieee.org/abstract/document/8518423" target="_blank">Urban-rural boundary of Whuhan</a> <br> -->
<a> Urban-rural boundary of Whuhan</a> <br>
<a href="https://wiki.citius.usc.es/hiperespectral:sae-cd" target="_blank">Hermiston City area, Oregon </a><br>
<a href="https://rcdaudt.github.io/oscd" target="_blank">OSCD </a><br>
<!-- <a href="https://ieeexplore.ieee.org/abstract/document/8519440" target="_blank">Quasi-urban areas </a><br> -->
<!-- <a > Quasi-urban areas </a><br> -->
<a href="https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html" target="_blank">WHU building dataset</a><br>
<a href="https://drive.google.com/file/d/1GX656JqqOyBi_Ef0w65kDGVto-nHrNs9/edit" target="_blank">Season-varing Dataset </a><br>
<a href="https://github.com/gistairc/ABCDdataset" target="_blank">ABCD </a><br>
<a href="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8798991" target="_blank"> California flood dataset </a><br>
<a href="https://citius.usc.es/investigacion/datasets/hyperspectral-change-detection-dataset" target="_blank"> Lopez-Fandino et al.</a><br>
<a href="https://ieee-dataport.org/open-access/hrscd-high-resolution-semantic-change-detection-dataset" target="_blank"> HRSCD </a> <br>
<a href="https://xview2.org/dataset" target="_blank"> xBD </a><br>
<a href="https://justchenhao.github.io/LEVIR" target="_blank"> LEVIR-CD </a> <br>
<a href="http://www.captain-whu.com/project/SCD/" target="_blank">SECOND </a> <br>
<a href="https://github.com/daifeng2016/Change-Detection-Dataset-for-High-Resolution-Satellite-Imagery" target="_blank"> Google Dataset </a><br>
<a href="https://github.com/MinZHANG-WHU/FDCNN" target="_blank"> Zhang et al.</a><br>
<a href="https://arxiv.org/abs/2011.03247" target="_blank"> Hi-UCD </a><br>
<a href="https://spacenet.ai/sn7-challenge/" target="_blank"> SpaceNet7 </a><br>
<a href="https://zenodo.org/record/4280482#.YDd3v-gzZPY" target="_blank"> S2MTCP </a><br>
</td>
<td style="text-align: center;">2 <br> 2 <br> 4 <br> 3 <br> 2 <br> 9 <br> 4 <br> 2 <br> 20 <br> 5 <br> 2
<br> 2 <br> 2 <br> 2 <br> 2 <br> 5 <br> 6 <br> 6 <br> 2 <br> 30 <br> 2 <br>2 <br> 9 <br> -- <br> 2 </td>
<td style="text-align: center;">13<br> 1,000 <br> 1 <br> 1 <br> 4 <br> 1 <br> 2 <br> 1 <br> 1 <br> 1 <br> 24 <br>
1 <br> 16,000 <br> 16,950 <br> 1 <br> 2 <br> 291 <br> 11034 <br> 637 <br> 4,214 <br> 1067 <br> 4 <br> 1,293 <br> 24 <br> 1,520 </td>
<td style="text-align: center;">1.5 <br> 0.5 <br> 30 <br> 30 <br>30, 120 <br> 1 <br> 30 <br> 30 <br> 4/30 <br> 30
<br> 10 <br> 0.2 <br> 0.03 to 0.1 <br> 0.4 <br> 5,30 <br> 20 <br> up to 0.8 <br> 0.5 <br> 0.5 <br>
0.5 to 3 <br> 0.55 <br> 2,2.4,5.8 <br> 0.1 <br> 4 <br> up to 10 </td>
<td style="text-align: center;">RGB <br> 115 bands <br> 6 bands <br> 6 bands <br> 7,9 <br> NIR, RGB <br> 242 bands <br>198
<br> 4/9 bands <br> 242 bands<br> 13 bands <br> RGB <br> RGB <br> RGB <br> RGB,11 <br> 224 <br>
RGB <br> RGB <br> RGB <br> RGB <br> RGB <br> NIR,RGB <br> RGB <br> RGB <br> 13 </td>
<td style="text-align: center;">952×640 <br> 800×600 <br> 400×400 <br> 800×800 <br>
444×300;1534×808<br> 7,200×6,000 <br> 400×145 <br> 463×241<br> 960×960 <br>
390×200 <br> 600×600 <br> 32,207×15,354 <br> 256×256 <br>
128×128;160×160 <br> 1534×808<br> 984×740;600×500<br> 10,000×10,000 <br> 1024×1,024<br>
1,024×1,024 <br> 512×512 <br> 256×256<br>1,431×1,431;458×559;1,154×740 <br> 1,024×1,024 <br> -- <br> 600×600 </td>
<td style="text-align: center;">✗✓✗ <br> ✗✗✗ <br> ✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✓✓ <br> ✓✓✓ <br>
✓✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✓✗ <br> ✓✓✓ <br> ✗✓✓ <br> ✓✓✓ <br> ✓✓✓ <br> ✗✗✗ <br> ✗✗✗ <br>
✓✓✗ <br> ✓✓✓ <br> --/--/✓ <br> ✓✓✓ <br> ✓✓✓ </td>
<td style="text-align: center;">2009 <br> 2011 <br> 2014 <br> 2014 <br> 2015 <br> 2017 <br> 2018 <br> 2018 <br> 2018 <br>
2018 <br> 2018 <br> 2018 <br> 2018 <br> 2018 <br> 2019 <br> 2019 <br> 2019 <br> 2019 <br> 2020 <br> 2020 <br> 2020<br>2020 <br> 2020 <br> 2020 <br> 2021</td>
</tr>
</tbody>
</table>
</div>
</div>
<br>
<div class="row">
<div class="span12">
<h3 id="DiRS" style="text-align:left; margin-bottom:10px; margin-top:20px; font-weight: bold">
3. DiRS: Principles to Build RS Image Benchmarks
</h3>
<p style="text-align:justify; font-size: 17px">
The primary point to construct a meaningful RS image dataset is that the dataset should be
created on the basis of the requirements of practical applications rather than the characteristics
of algorithms to be employed. Moreover, the annotation of RS image dataset is better to be conducted
by the application sides rather than the algorithm developers. Thus, the annotated dataset is
naturally application-oriented, which is more conducive to enhance the practicability of the
interpretation algorithm. With these points in mind, the <i>i.e.</i>, <i><strong>di</strong>versity</i>,
<i><strong>r</strong>ichness</i>, and <i><strong>s</strong>calability</i> (called <strong>DiRS</strong>)
could be considered as the basic principles when creating benchmark datasets for RS image interpretation.
We believe that these principles are complementary to each other. That is, the improvement of dataset in
one principle can simultaneously promote the dataset quality reflected in other principles.
</p>
<img src="files/DiRS.jpg" width="750px" class="img-responsive center-block" />
</div>
</div>
<br>
<div class="row">
<div class="span12">
<h3 id="Million-AID" style="text-align:left; margin-bottom:10px; margin-top:20px; font-weight: bold">
4. An Example: Million-AID
</h3>
<a href="files/Million-AID.jpg" target="_blank" title="Click to enlarge">
<img src="files/Million-AID-Head.jpg" style="margin-right: 20px;margin-top: 5px; float: left;width: 420px;"/>
</a>
<p style="text-align:justify; font-size: 17px;">
Following the DiRS principles, we provide an example on building datasets for RS image classification,
<i>i.e.</i>, Million-AID, a new large-scale benchmark dataset containing million instances for RS scene
classification. Million-AID contains a wide range of semantic categories, <i>i.e.</i>, 51 scene categories
organized by the hierarchical category network of a three-level tree: 51 leaf nodes fall into 28 parent
nodes at the second level which are grouped into 8 nodes at the first level, representing the 8 underlying
scene categories of agriculture land, commercial land, industrial land, public service land, residential
land, transportation land, unutilized land, and water area. The scene category network provides the dataset
with excellent organization of relationship among different scene categories and also the property of
scalability. The number of images in each scene category ranges from about 2,000 to 45,000, endowing the
dataset with the property of long tail distribution. Besides, Million-AID has superiorities over the
existing scene classification datasets owing to its high spatial resolution, large scale, and global
distribution.
</p>
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Category Network
</h4>
<img src="files/MAID_ClsNet.jpg" width="750px" class="img-responsive center-block" />
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Semantic Coordinates Collection
</h4>
<img src="files/CoordinatesCollection.jpg" width="750px" class="img-responsive center-block" />
<h4 style="text-align:left; margin-bottom:10px; margin-top:10px; font-weight: bold; font-style: italic">
- Scene Image Acquisition
</h4>
<img src="files/SceneAcquisition.jpg" width="750px" class="img-responsive center-block" />
<h3 id="downeva" style="text-align:left; margin-bottom:10px; margin-top:30px; font-weight: bold">
Dataset & Evaluation
</h4>
<p style="text-align:justify; font-size: 17px">
Million-AID has been released for public accessibility.
</p>
<ul>
<li style="font-size:17px">
<!-- <a href="">Million-AID Download (OneDrive)</a> (<i>Coming soon ...</i> ) -->
<a href="https://whueducn-my.sharepoint.com/:f:/g/personal/longyang_whu_edu_cn/Et-SJsQYQRxMh63Z59iFyH0BramZuLnyo4XKoi5yrbfb9A" target="_blank">Million-AID Download (One Drive)</a>
<a href="https://pan.baidu.com/s/1URt_dyAybExu9fsOg4lZgQ" target="_blank">Million-AID Download (Baidu Drive, extraction code: 107t)</a>
</li>
<li style="font-size:17px">
<a href="https://captain-whu.github.io/ASP/#scbm" target="_blank">Benchmarking Results</a>
</li>
<li style="font-size:17px">
<a href="https://competitions.codalab.org/competitions/35945" target="_blank">Evaluation Server for Multi-class Scene Classification</a>
</li>
<li style="font-size:17px">
<a href="https://competitions.codalab.org/competitions/35974" target="_blank">Evaluation Server for Multi-label Scene Classification</a>
</li>
</ul>
<!-- <a href="files/Million-AID.jpg" target="_blank" title="Click to enlarge"><img src="files/Million-AID-Display.jpg" width="750px" class="img-responsive center-block" /></a> -->
</div>
</div>
<br>
<div class="row">
<div class="span12">
<div class="section bibtex">
<h3 style="text-align:left; margin-bottom:10px; margin-top:20px; font-weight: bold">
Citation
</h3>
<p> If you want to make use of Million-AID, please cite our following paper:</p>
<pre>
@article{Long2021DiRS,
title={On Creating Benchmark Dataset for Aerial Image Interpretation: Reviews, Guidances and Million-AID},
author={Yang Long and Gui-Song Xia and Shengyang Li and Wen Yang and Michael Ying Yang and Xiao Xiang Zhu and Liangpei Zhang and Deren Li},
journal={IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
year={2021},
volume={14},
pages={4205-4230}
}</pre>
<br>
<pre>
@misc{Long2022ASP,
title={Aerial Scene Parsing: From Tile-level Scene Classification to Pixel-wise Semantic Labeling},
author={Yang Long and Gui-Song Xia and Liangpei Zhang and Gong Cheng and Deren Li},
year={2022},
eprint={2201.01953},
archivePrefix={arXiv},
primaryClass={cs.CV}
}</pre>
</div>
<h3 style="text-align:left; margin-bottom:10px; margin-top:30px; font-weight: bold">
Contact
</h3>
<p>
If you have any problem, please contact:
<ul>
<li>Yang Long at <strong>[email protected]</strong></li>
<!-- <li>Gui-Song Xia at <strong>[email protected]</strong></li> -->
</ul>
<br />
<br />
<br />
</div>
</div>
<!-- <div class="row">
<div style="text-align:center; margin-top:0; margin-bottom: 20px;">
<embed id="map" src="http://rf.revolvermaps.com/f/f.swf" type="application/x-shockwave-flash" pluginspage="http://www.macromedia.com/go/getflashplayer" wmode="transparent" allowScriptAccess="always" allowNetworking="all" width="150" height="75" flashvars="m=0&i=5dp1mfnunae&r=10&c=fffdc0" loop="true" autostart="False"></embed>
<img class="img-responsive center-block" src="http://rf.revolvermaps.com/js/c/5dp1mfnunae.gif" width="1" height="1" alt="" value="True"/>
<a style="font-size: x-small;"> Copyight@2020, Captain</a>
<a href="http://www.revolvermaps.com/livestats/5dp1mfnunae/"></a>
</div>
</div> -->
</body>
</html>