Optimum Fusion Method and Application of Landslide Remote Sensing in Mountainous Woodland Areas

doi:10.11873/j.issn.1004-0323.2022.5.1248

Remote Sensing Technology and Application

2022, Vol. 37

Issue (5): 1248-1258 DOI: 10.11873/j.issn.1004-0323.2022.5.1248

Optimum Fusion Method and Application of Landslide Remote Sensing in Mountainous Woodland Areas

Yuqing Shi^1,²(

),Ji Liang^1,²(

),Yunxing Li³,Saiying Meng^1,²,Qian Shi^1,²

^1.National-Local Joint Engineering Laboratory of Geo-Spatial Information Technology in the Hunan University of Science and Technology，Xiangtan 411201，China
^2.School of Geosciences and Spatial Information Engineering，Hunan University of Science and Technology，Xiangtan 411201，China
^3.the First Surveying and Mapping Institute of Hunan Province，Changsha 410118，China

Download:

HTML

PDF (5150KB)
Export: BibTeX | EndNote (RIS)

Abstract

As one of the geological disasters causing huge economic losses and casualties， landslides have attracted more and more attention from society. In order to accurately identify landslide disasters in mountainous woodland areas， the Leijiashan landslide， which occurred on July 6， 2020 in Panping Village， Nanbei Town， Shimen County， Changde City， Hunan Province， was taken as the research object. Different fusion methods such as Principal Component Analysis （PCA）， Gram-Schmidt （GS） and Nearest-Neighbor Diffusion （NNDiffuse） are used to fuse the images of Sentinel-1A Interferometric Wide Swath （IW） Ground Range Detected （GRD） image after non-decibelization and decibelization with Sentinel-2A MSI2A image. Through the quality evaluation of the fused image， the PCA fusion method effect of the VV polarization image of Sentinel-1A after decibelization and Sentinel-2A image is the optimal， that is， the optimal fusion image is PCA-VV-DB. The Support Vector Machine （SVM） method was used to identify the landslide of the optimal fusion image （PCA-VV-DB） and the original optical image Sentinel-2A， respectively. Finally， the Sentinel-2A landslide visual interpretation results were used as the inspection standard to evaluate and compare the accuracy of SVM landslide identification results. At the same time， the Shaziba landslide in Mazhe Village， Tunbao Township， Enshi City， Hubei Province， on July 21， 2020， was used as a case to verify the feasibility of this scheme. The results show that compared with the single use of optical image for landslide recognition in the study area， the accuracy of landslide recognition using the optimal fusion image is increased from 95.24% to 96.65%， and the quality of landslide extraction also increased from 87.18% to 91.84%. The leakage recognition and excessive recognition of landslides are reduced， and the research scheme is popularized. It shows that the fusion of optical image and Synthetic Aperture Radar （SAR） image can improve the accuracy of landslide recognition in mountainous woodland areas， and provide valuable information for landslide risk assessment， disaster emergency investigation and disaster recovery and reconstruction.

Key words: Mountainous woodland areas Landslide identification Optical image SAR image Fusion SVM

Received: 29 July 2021 Published: 13 December 2022

ZTFLH:

TP753

Corresponding Authors: Ji Liang E-mail: 826583940@qq.com;leung@lzb.ac.cn

	Service
	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Yuqing Shi
	Ji Liang
	Yunxing Li
	Saiying Meng
	Qian Shi

Cite this article:

Yuqing Shi,Ji Liang,Yunxing Li,Saiying Meng,Qian Shi. Optimum Fusion Method and Application of Landslide Remote Sensing in Mountainous Woodland Areas. Remote Sensing Technology and Application, 2022, 37(5): 1248-1258.

URL:

http://www.rsta.ac.cn/EN/10.11873/j.issn.1004-0323.2022.5.1248 OR http://www.rsta.ac.cn/EN/Y2022/V37/I5/1248

Fig.1 Location of study area

Fig.2 Historical weather data from January to July in Shimen County

Table 1 Sentinel-1A and Sentinel-2A data parameters

Fig.3 Flow chart of landslide identification technology

评价指标	计算公式	参数及含义
均值（ $A ˉ$ ），表示影像灰度的平均值。理想的融合影像应适当增大影像的均值	$A ˉ = 1 M N ∑ i = 1 M ∑ j = 1 N A (i, j)$	M和N分别为影像的宽度和高度； A(i,j)为影像的灰度值
标准差（SD），反映影像各像元灰度的离散情况。标准差越大，空间信息越丰富，融合影像的效果越好	$S D = 1 M N ∑ i = 1 M ∑ j = 1 N (A i, j - A ˉ) 2$	$A ˉ$ 为均值
信息熵（H），衡量影像的信息丰富程度。融合图像的信息熵越大，说明融合影像的信息量增加的越多	$H = - ∑ i = 1 L P i ? l o g 2 ? P i$	P_i 为影像中像素灰度值为i的概率； L为影像的总灰度级数
平均梯度（ $G ˉ$ ），反映影像对微小细节的表达能力。平均梯度越大，影像的清晰度越高	$G ˉ = 1 M N ∑ i = 1 M ∑ j = 1 N Δ I x 2 + Δ I y 2 ∕ 2$	ΔIx和ΔIy分别为x和y方向上的一阶差分
相关系数（R），反映两幅影像的相关程度。相关系数越大，融合方法越能保持原始影像的光谱特征	$R = ∑ i = 1 M ∑ j = 1 N (A i, j - A ˉ) (F i, j - F ˉ) ∑ i = 1 M ∑ j = 1 N [A i, j - A ˉ] 2 × ∑ i = 1 M ∑ j = 1 N [F i, j - F ˉ] 2$	F(i,j)为融合影像的灰度值； $F ˉ$ 为融合影像的均值
扭曲程度（D），用于评价多光谱的信息的保持程度。扭曲程度越小，表明影像的失真程度越小	$D = 1 M N ∑ i = 1 M ∑ j = 1 N F i, j - A i, j$
峰值信噪比（PSNR），衡量影像失真或噪声水平的客观指标。峰值信噪比越大，说明融合效果和质量越好	$P S N R = 10 l g ? M N ? m a x F ?, j 2 ∑ i = 1 M ∑ j = 1 N A i, j - F i, j 2$	max[F(i,j)]为融合影像最大灰度值

Table 2 Evaluation index

Fig.4 Fusion results of S1A GRD and S2A MSI2A images

数据	波段	均值	标准差	信息熵	平均梯度	相关系数	扭曲程度	峰值信噪比
Sentinel-2A	B2	258.65	178.67	1.27	56.43	1.00	——	——
	B3	442.20	241.09	1.74	78.05	1.00	——	——
	B4	417.30	306.18	1.70	82.10	1.00	——	——
	B8	2 109.70	796.63	3.60	312.91	1.00	——	——
PCA-VH	B2	258.72	172.50	1.31	54.92	0.78	84.49	55.05
	B3	442.24	229.41	1.66	71.37	0.59	154.57	49.80
	B4	417.50	296.06	1.84	82.22	0.81	136.59	50.88
	B8	2 109.70	805.04	2.94	168.39	-0.04	841.45	35.07
PCA-VV	B2	258.70	175.87	1.31	55.11	0.79	82.75	55.08
	B3	442.22	235.24	1.65	71.88	0.61	151.38	49.83
	B4	417.48	302.14	1.83	82.52	0.81	133.78	50.92
	B8	2 109.70	800.39	2.81	170.31	-0.04	824.03	35.11
PCA-VH-DB	B2	258.78	171.05	1.41	55.65	0.78	94.56	54.98
	B3	442.25	226.99	1.86	73.88	0.58	173.09	49.72
	B4	417.85	292.44	2.02	83.38	0.80	152.64	50.83
	B8	2 109.70	807.14	3.67	218.32	-0.06	942.35	35.00
PCA-VV-DB	B2	259.06	177.96	1.44	55.63	0.79	91.70	55.10
	B3	442.79	238.98	1.89	74.03	0.61	167.83	49.84
	B4	418.33	305.11	2.03	83.55	0.82	147.98	50.95
	B8	2 110.90	793.18	3.65	218.37	-0.05	915.43	35.10
GS-VH	B2	260.23	173.36	1.46	57.31	0.24	154.47	49.62
	B3	443.41	233.78	1.70	69.44	0.08	231.64	46.19
	B4	421.05	295.50	2.09	89.45	0.31	254.13	45.36
	B8	2 105.70	815.55	3.35	223.49	0.15	771.01	35.91
GS-VV	B2	260.18	176.60	1.46	57.99	0.27	151.07	49.72
	B3	443.39	237.75	1.70	70.35	0.12	225.70	46.30
	B4	420.90	302.31	2.07	90.49	0.34	248.17	45.46
	B8	2 105.60	801.37	3.29	223.99	0.16	750.83	36.02
GS-VH-DB	B2	262.53	169.04	1.57	58.26	0.23	171.17	49.64
	B3	444.20	231.73	1.96	75.00	0.06	260.16	46.12
	B4	426.77	283.36	2.19	90.01	0.29	280.61	45.42
	B8	2 108.80	809.65	3.68	243.63	0.14	853.84	35.87
GS-VV-DB	B2	263.79	173.60	1.60	58.41	0.30	164.52	49.93
	B3	446.30	236.19	1.98	74.70	0.15	248.79	46.44
	B4	428.69	294.43	2.23	90.65	0.37	269.48	45.73
	B8	2 106.70	782.13	3.64	244.60	0.15	823.00	36.08
NNDiffuse-VH	B2	264.35	276.06	1.25	62.32	0.64	13.38	49.78
	B3	449.61	327.00	1.74	85.52	0.70	17.28	49.01
	B4	425.99	408.72	1.70	93.17	0.73	18.66	47.45
	B8	2 160.50	1599.60	3.59	391.22	0.47	86.78	33.34
NNDiffuse-VV	B2	281.77	439.48	1.30	98.22	0.41	30.61	44.28
	B3	475.45	602.57	1.79	138.42	0.38	42.92	41.39
	B4	458.34	801.01	1.75	159.91	0.40	50.75	39.01
	B8	2 284.40	2646.40	3.64	641.98	0.26	209.96	28.16
NNDiffuse-VH-DB	B2	258.15	175.94	1.25	52.04	1.00	4.10	79.32
	B3	441.70	237.80	1.73	72.59	1.00	5.47	77.48
	B4	416.80	302.81	1.69	76.90	1.00	5.58	76.71
	B8	2 109.20	782.83	3.58	293.23	1.00	21.03	66.54
NNDiffuse-VV-DB	B2	258.15	175.86	1.25	51.90	1.00	4.22	79.08
	B3	441.71	237.76	1.73	72.46	1.00	5.56	77.38
	B4	416.80	302.74	1.69	76.76	1.00	5.68	76.58
	B8	2 109.20	782.83	3.58	292.96	1.00	21.15	66.56

Table 3 Evaluation results of fusion image quality

Fig.5 Results of artificial visual interpretation of landslide in S2A image

Fig.6 SVM landslide recognition results of S2A image and S1A and S2A optimal fusion image

Table 4 Accuracy of landslide identification

Table 5 Evaluation results of fusion image quality

Fig.7 SVM landslide recognition results of S2A image and S1A and S2A optimal fusion image

Table 6 Accuracy of landslide identification

1	Liao Mingsheng， Zhang Lu， Shi Xuguo， et al. Methodology and practice of landslide deformation monitoring with SAR remote sensing［M］. Beijing：Science Press， 2017.廖明生，张路，史绪国，等.滑坡变形雷达遥感监测方法与实践［M］.北京：科学出版社， 2017.
2	Chen Wenlong， Hou Yong， Li Nan， et al. Post-earthquake landslide detection in nepal based on principal component analysis［J］. Journal of Yangtze River Scientific Research Institute， 2020， 37（1）： 166-171.
2	陈文龙，候勇，李楠，等.基于主成分变换的滑坡识别方法及其在2015年尼泊尔地震中的应用［J］.长江科学院院报， 2020， 37（1）： 166-171.
3	Peng Ling， Xu Suning， Mei Junjun， et al. Earthquake-induced landslide recognition using high-resolution remote sensing images［J］. Journal of Remote Sensing，2017，21（4）： 509-518.
3	彭令，徐素宁，梅军军，等.地震滑坡高分辨率遥感影像识别［J］.遥感学报， 2017， 21（4）： 509-518.
4	Borghuis A M， Chang K， Lee H Y. Comparison between automated and manual mapping of typhoon-triggered landslides from SPOT5 imagery［J］. International Journal of Remote Sensing，2007，28（7/8）：1843-1856. DOI： . doi: 10.1080/014311 60600935638
5	Zhao W， Li A N， Nan X， et al. Postearthquake landslides mapping from Landsat-8 data for the 2015 Nepal Earthquake using a pixel-based change detection method［J］. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing，2017，10（5）：1758-1768. DOI： . doi: 10.1109/jstar. 2017.2661802
6	Deng Shubin， Chen Qiujin， Du Huijian， et al. ENVI remote sensing image processing.Second Edition［M］.Beijing：Higher Education Press， 2014.
6	邓书斌，陈秋锦，杜会建，等.ENVI遥感图像处理方法.第2版［M］.北京：高等教育出版社，2014.
7	Fu Wenjie， Hong Jinyi. Discussion on application of support vector machine technique in extraction of information on landslide hazard from remote sensing images［J］. Research of Soil and Water Conservation， 2006， 13（4）： 120-124.
7	傅文杰，洪金益.基于支持向量机的滑坡灾害信息遥感图像提取研究［J］.水土保持研究， 2006， 13（4）：120-124.
8	Marjanovic M， Bajat B， Kovacevic M. Landslide susceptibility assessment with machine learning algorithms［C］// International Conference on Intelligent Networking & Collaborative Systems. IEEE， 2009. DOI： . doi: 10.1109/incos.2009.25
9	Dai Cong， Li Weile， Lu Huiyan， et al. Active landslides detection in Zhouqu County， Gansu Province using InSAR technology［J］. Geomatics and Information Science of Wuhan University， 2021， 46（7）： 994-1002.
9	代聪，李为乐，陆会燕，等.甘肃省舟曲县城周边活动滑坡InSAR探测［J］.武汉大学学报（信息科学版）， 2021， 46（7）： 994-1002.
10	Dai Keren， Yongbo Tie， Xu Qiang， et al. Early identification of potential landslide geohazards in Alpine-canyon terrain based on SAR interferometry： A case study of the middle section of Yalong River［J］. Journal of Radars，2020，9（3）：554-568.
10	戴可人，铁永波，许强，等.高山峡谷区滑坡灾害隐患InSAR早期识别——以雅砻江中段为例［J］.雷达学报， 2020， 9（3）： 554-568.
11	Dai Keren， Zhuo Guanchen， Xu Qiang， et al. Tracing the pre-failure two-dimensional surface displacements of Nanyu landslide， Gansu Province with radar interferometry［J］. Geomatics and Information Science of Wuhan University， 2019， 44（12）： 1778-1786，1796.
11	戴可人，卓冠晨，许强，等.雷达干涉测量对甘肃南峪乡滑坡灾前二维形变追溯［J］.武汉大学学报（信息科学版）， 2019， 44（12）：1778-1786，1796.
12	Jiang Mi， Ding Xiaoli， Li Zhiwei， et al. Study on coseismic deformation of Wenchuan Earthquake by USE of land C wavebands of SAR data［J］. Journal of Geodesy and Geodynamics， 2009， 29（1）：21-26.
12	蒋弥，丁晓利，李志伟，等.用L波段和C波段SAR数据研究汶川地震的同震形变［J］.大地测量与地球动力学， 2009， 29（1）： 21-26.
13	Wang Liwen， Wei Yaxing. Progress in monitoring wetland ecosystems by radar remote sensing［J］. Progress in Geography，2011，30（9）： 1107-1117.
13	王莉雯，卫亚星.湿地生态系统雷达遥感监测研究进展［J］.地理科学进展， 2011， 30（9）： 1107-1117.
14	Wang Xuan， Fan Xuanmei， Yang Fan， et al. Remote sensing interpretation method of geological hazards in Lush Mountainous area［J］. Geomatics and Information Science of Wuhan University， 2020， 45（11）： 1771-1781.
14	王绚，范宣梅，杨帆，等.植被茂密山区地质灾害遥感解译方法研究［J］.武汉大学学报（信息科学版）， 2020， 45（11）： 1771-1781.
15	Guo Jiao， Zhu Lin， Jin Biao. Crop classification based on data fusion of Sentinel-1 and Sentinel-2［J］. Transactions of The Chinese Society of Agricultural Machinery，2018，49（4）：192-198.
15	郭交，朱琳，靳标.基于Sentinel-1和Sentinel-2数据融合的农作物分类［J］.农业机械学报，2018，49（4）：192-198.
16	Gaetano R， Cozzolino D， D'Amiano L， et al. Fusion of sar-optical data for land cover monitoring［C］// IGARSS 2017 - 2017 IEEE International Geoscience and Remote Sensing Symposium.IEEE，2017. DOI： . doi: 10.1109/igarss.2017.8128242
17	Gargiulo M， Dell'Aglio D A G， Iodice A， et al. Integration of Sentinel-1 and Sentinel-2 data for land cover mapping using W-Net［J］.Sensors，2020，20（10）：2969. DOI： . doi: 10.3390/s 20102969
18	Chen Jing， Yanjun Ou. The supremacy of the people and the supremacy of life： A profile of the response to landslides in the Nanbei Towns of Shimen County［J］. Hunan Security and Disaster Prevention， 2020（8）： 28-29.
18	陈晶，欧彦君.人民至上生命至上——石门县南北镇应对山体滑坡灾害侧记［J］.湖南安全与防灾， 2020（8）： 28-29.
19	Liang Ji， Chu Nan， Zheng Dunyong， et al. Application to GF-2 satellite for road landslides monitoring［J］.Remote Sensing Technology and Application， 2018， 33（4）： 638-645.
19	梁继，褚楠，郑敦勇，等.面向道路边坡监测的高分二号应用研究［J］.遥感技术与应用， 2018， 33（4）： 638-645.
20	Sanli F B， Abdikan S， Esetlili M T， et al. Evaluation of image fusion methods using PALSAR， Radarsat-1 and SPOT images for land use/land cover classification［J］. Journal of the Indian Society of Remote Sensing， 2016： 1-11. DOI： . doi: 10.1007/s12524-016-0625-y
21	Liu Kun， Fu Jingying， Li Fei. Evaluation study of four fusion methods of GF-1 PAN and multi-spectral images［J］. Remote Sensing Technology and Application， 2015，30（5）：980-986.
21	刘锟，付晶莹，李飞.高分一号卫星4种融合方法评价［J］.遥感技术与应用， 2015，30（5）： 980-986.
22	Qin Shanshan， Wang Shixin， Zhou Yi， et al. A research on fusion method for ZY-3 satellite data based on NSCT and GS transform［J］. Journal of Geo-Information Science，2014，16（6）： 949-957.
22	秦善善，王世新，周艺，等.NSCT与GS变换的资源三号卫星数据融合方法研究与应用［J］.地球信息科学学报， 2014， 16（6）： 949-957.
23	Sun W， Chen B， Messinger D W. Nearest-neighbor diffusion-based pan-sharpening algorithm for spectral images［J］. Optical Engineering， 2013， 53（1）： 013107. DOI： . doi: 10.1117/1.oe.53.1.013107
24	Zhao J， Guo J， Cheng W， et al. Assessment of SPOT-6 optical remote sensing data against GF-1 using NNDiffuse image fusion algorithm［J］. Modern Physics Letters B， 2017，31（19-21）：1740043. DOI： . doi: 10.1142/s0217984917400437
25	Xu Jing， An Yulun， Liu Suihua， et al. Image fusion of spaceborne SAR data and multi spectral data for mountainous plateau： A case study on Bijie City， Guizhou Province， China［J］. Earth and Environment， 2015， 43（4）：457-463.
25	许璟，安裕伦，刘绥华，等.高原山区星载合成孔径雷达数据与多光谱数据的图像融合探究——以贵州省毕节市为例［J］.地球与环境， 2015， 43（4）： 457-463.
26	Wang Fang， Yang Wunian， Wang Jian， et al. Chinese high-resolution satellite pixel level image fusion and its quality evaluation［J］. Science of Surveying and Mapping， 2021， 46（8）： 73-80.
26	王芳，杨武年，王建，等.国产高分卫星像素级影像融合及其质量评价［J］.测绘科学， 2021， 46（8）： 73-80.
27	Yang Jun， Wang Xiaoyu. Comparison research on fusion methods of GF-2 and Sentinel-2 Panchromatic multispectral images［J］. Science of Surveying and Mapping，2022，47（1）： 112-120.
27	杨军，王筱宇.GF-2和Sentinel-2全色多光谱影像融合方法比较研究［J］.测绘科学， 2022， 47（1）：112-120.
28	Abdikan Saygin. Exploring image fusion of ALOS/PALSAR data and LANDSAT data to differentiate forest area［J］. Geocarto International，2016，33（1）：21-37. DOI： . doi: 10.1080/10106049. 2016.1222635
29	Yang Liping， Ma Meng， Xie Wei， et al. Fusion algorithm evaluation of Landsat 8 panchromatic and multispetral images in arid regions［J］. Remote Sensing for Land and Resources， 2019， 31（4）：11-19.
29	杨丽萍，马孟，谢巍，等.干旱区Landsat8全色与多光谱数据融合算法评价［J］.国土资源遥感， 2019， 31（4）： 11-19.
30	Heqianshui Super. SNAP Processes Sentinel-1 IW GRD Data［EB/OL］. ， 2020.
30	超级禾欠水.SNAP处理Sentinel-1 IW GRD数据［EB/OL］. ， 2020.
31	Gong P， Liu H， Zhang M N， et al. Stable classification with limited sample： Transferring a 30 m resolution sample set collected in 2015 to mapping 10 m resolution global Land Cover in 2017［J］. Science Bulletin， 2019， 64（6）： 370-373. DOI： . doi: 10.1016/j.scib.2019.03.002
32	Lin Qigen， Zou Zhenhua， Zhu Yingqi， et al. Object-oriented detection of landslides based on the spectral， spatial and morphometric properties of landslides［J］. Remote Sensing Technology and Application， 2017， 32（5）：931-9.
32	林齐根，邹振华，祝瑛琦，等.基于光谱、空间和形态特征的面向对象滑坡识别［J］.遥感技术与应用， 2017， 32（5）： 931-937.

No Suggested Reading articles found!

Viewed

Full text

Abstract

Cited

Shared

Discussed