-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
106 lines (88 loc) · 6.76 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
<!-- saved from url=(0042)https://x-ytong.github.io/project/GID.html -->
<html lang="en"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>GID Dataset</title>
<style id="system" type="text/css">
h1,h2,h3,h4,h5,h6,p,blockquote { margin: 0; padding: 0;}
body {font-family: "Helvetica Neue", Helvetica, "Hiragino Sans GB", Arial, sans-serif;
font-size: 15px; line-height: 20px; color: #737373; margin: 10px 13px 10px 13px;}
tr,td {font-size: 15px; color: #737373; text-align: justify; text-justify: inter-character;}
ul {text-align: justify; text-justify: inter-character;}
a {color: #ff7f50; text-decoration: none;}
a:hover {color: #c39797; text-decoration: none;}
a img {border: none;}
p {margin-bottom: 9px; text-align: justify; text-justify: inter-character;}
h1,h2,h3,h4,h5,h6 {color: #20b2aa; line-height: 36px;}
h1 {margin-top: 50px; margin-bottom: 18px; font-size: 26px;}
h2 {margin-top: 24px; font-size: 20px;}
h3 {font-size: 18px;}
h4 {font-size: 16px;}
h5 {font-size: 14px;}
h6 {font-size: 13px;}
hr {margin: 0 0 19px; border: 0; border-bottom: 1px solid #ccc;}
blockquote {padding: 13px 13px 21px 15px; margin-bottom: 18px; font-family:georgia,serif; font-style: italic;}
blockquote:before {content:"C"; font-size:40px; margin-left:-10px; font-family:georgia,serif; color:#eee;}
blockquote p {font-size: 14px; font-weight: 300; line-height: 18px; margin-bottom: 0; font-style: italic;}
code, pre {font-family: Monaco, Andale Mono, Courier New, monospace;}
code {background-color: #fee9cc; color: rgba(0, 0, 0, 0.75); padding: 1px 3px; font-size: 12px;
-webkit-border-radius: 3px; -moz-border-radius: 3px; border-radius: 3px;}
pre {display: block; padding: 14px; margin: 0 0 18px; line-height: 16px; font-size: 11px;
border: 1px solid #d9d9d9; white-space: pre-wrap; word-wrap: break-word;}
pre code {background-color: #fff; color:#737373; font-size: 11px; padding: 0;}
@media screen and (min-width: 768px) {body {width: 748px; margin:10px auto;}}
</style><style id="custom" type="text/css"></style></head>
<body marginheight="0">
<div align="center"><h1>Land-Cover Classification with High-Resolution Remote Sensing Images Using Transferable Deep Models</h1></div>
<div style="border:2px solid #FFFFFF "></div>
<div align="center">Xin-Yi Tong, Gui-Song Xia, Qikai Lu, Huanfeng Shen, Shengyang Li, Shucheng You, Liangpei Zhang</div>
<h2>Abstract</h2>
<p>In recent years, large amount of high spatial-resolution remote sensing (HRRS) images are available for land-cover mapping.
However, due to the complex information brought by the increased spatial resolution and the data disturbances caused by
different conditions of image acquisition, it is often difficult to find an efficient method for achieving accurate land-cover
classification with high-resolution and heterogeneous remote sensing images. In this paper, we propose a scheme to apply deep
model obtained from labeled land-cover dataset to classify unlabeled HRRS images. The main idea is to rely on deep neural
networks for presenting the contextual information contained in different types of land-covers and propose a pseudo-labeling
and sample selection scheme for improving the transferability of deep models. To pre-train deep model specific to HRRS images,
we create a large-scale land-cover dataset containing <b>150 Gaofen-2 satellite images</b>. Experiments on multi-source HRRS
images, including Gaofen-2, Gaofen-1, Jilin-1, Ziyuan-3, Sentinel-2A, and Google Earth platform data, show encouraging results
and demonstrate the applicability of the proposed scheme to land-cover classification with multi-source HRRS images.</p>
<h2>Dataset</h2>
<p>We construct a large-scale land-cover dataset with Gaofen-2 (GF-2) satellite images. This new dataset, which is named as
Gaofen Image Dataset (GID), has superiorities over the existing land-cover dataset because of its large coverage, wide distribution,
and high spatial resolution. GID consists of two parts: a large-scale classification set and a fine land-cover classification set.
The large-scale classification set contains 150 pixel-level annotated GF-2 images, and the fine classification set is composed of
30,000 multi-scale image patches coupled with 10 pixel-level annotated GF-2 images. The training and validation data with 15
categories is collected and re-labeled based on the training and validation images with 5 categories, respectively.</p>
<h3>- Large-scale classification set</h3>
<div align="center"><img src="./GID Dataset_files/GIDlarge.png" width="70%"></div>
<h3>- Fine land-cover classification set</h3>
<div align="center"><img src="./GID Dataset_files/GIDfine.png" width="70%"></div>
<h3>- Multi-source validation images</h3>
<div align="center"><img src="./GID Dataset_files/GIDmulti.png" width="70%"></div>
<h3>- Download</h3>
<p>GID can be download from Onedrive or Baidudrive:</p>
<p></p><li>Link: <a href="https://whueducn-my.sharepoint.com/:f:/g/personal/xinyi_tong_whu_edu_cn/El2LVTJPk3VAuL7_eENrZ68BLPDsJJvm80uYbwXABx5Wuw?e=GwePwA/"><b>Onedrive</b></a></li><p></p>
<p></p><li>Link: <a href="https://whueducn-my.sharepoint.com/:f:/g/personal/xinyi_tong_whu_edu_cn/EmT3E8xDYaBHklxweG0Je68BGPRWaitir0CuSJaaCjJfDQ?e=87V6PA/"><b>Onedrive (compressed version)</b></a></li><p></p>
<p></p><li>Link: <a href="https://pan.baidu.com/s/1kdMdgXCUWFmlpaKXjFRaaA/"><b>Baidudrive</b></a> (extraction code:5r1z)</li><p></p>
<h2>Experiment</h2>
<p>We test our algorithm and analyse the experimental results in this section. Two types of land-cover classification issues are
examined: 1) transferring deep models to classify HRRS images captured with the same sensor and under different conditions,
2) transferring deep models to classify multi-source HRRS images. For performance comparison, several object-based land-cover
classification methods are utilized.</p>
<h3>- Experiments on Gaofen-2 images</h3>
<div align="center"><img src="./GID Dataset_files/GIDresultmap5.png" width="80%"></div>
<div style="border:10px solid #FFFFFF "></div>
<div align="center"><img src="./GID Dataset_files/GIDresultmap15.png" width="80%"></div>
<h3>- Experiments on multi-source images</h3>
<div align="center"><img src="./GID Dataset_files/GIDmultiresultmap15.png" width="80%"></div>
<h2>Citation</h2>
<pre>@article{GID2020,
title = {Land-Cover Classification with High-Resolution Remote Sensing Images Using Transferable Deep Models},
author = {Xin-Yi Tong, Gui-Song Xia, Qikai Lu, Huangfeng Shen, Shengyang Li, Shucheng You, Liangpei Zhang},
journal = {Remote Sensing of Environment, doi: 10.1016/j.rse.2019.111322},
year = {2020}
}
</pre>
<h2>Contact</h2>
<p>E-mail : [email protected]</p>
<div style="border:20px solid #FFFFFF "></div>
</body></html>