Welcome to Acta Agronomica Sinica,

Acta Agronomica Sinica ›› 2020, Vol. 46 ›› Issue (7): 1099-1111.doi: 10.3724/SP.J.1006.2020.94134

• TILLAGE & CULTIVATION·PHYSIOLOGY & BIOCHEMISTRY • Previous Articles     Next Articles

Extraction of crop acreage based on multi-temporal and dual-polarization SAR data

Gulnur ISAK,Mamat SAWUT(),MA Chun-Yue   

  1. College of Resources and Environmental Sciences, Xinjiang University / Ministry of Education Key Laboratory of Oasis Ecology, Xinjiang University / Key Laboratory for Wisdom City and Environmental Modeling, Xinjiang University, Urumqi 830046, Xinjiang, China
  • Received:2019-09-10 Accepted:2020-03-24 Online:2020-07-12 Published:2020-04-13
  • Contact: SAWUT Mamat E-mail:korxat@xju.edu.cn
  • Supported by:
    National Natural Science Foundation of China(41361016);National Natural Science Foundation of China(41461051)

Abstract:

It plays a vital role in agricultural production management and agricultural policy formulation to acquire spatial distribution information and planting area of crops timely and accurately. In this paper, according to the phenological characteristics of crops, the back scattering coefficients (Sigma) and normalized back scattering coefficients (Gamma) of cotton, corn and orchard in different growth periods were extracted by using multi-temporal Sentinel-1A images (April 17, May 5, June 16, July 22, August 27, and September 2). The time-series change curves of the scattering characteristics were established and the characteristics were analyzed by using backscattering coefficients of crops with different polarizations and different time phases. Artificial neural network (ANN), support vector machine (SVM), and random forest (RF) were used to classify and identify cotton, corn and orchard. On this basis, the classification results were compared and analyzed, and the accuracy was verified. The backscattering coefficient of cotton increased significantly in June and July with the highest value in August, of which the changes were most obvious and easily distinguished from other crops. The backscattering coefficients of corn and fruit trees were significantly different from those of other land objects in September. The optimal classification was obtained by using random forest (the overall accuracy was up to 88.97%) than by using neural network and support vector machine. The classification accuracy for cotton and orchard was 90.88% and 93.17%, and the classification effect of corn is the worst, only 71.6%. In conclusion, multi-temporal and dual-polarization SAR data has certain application potential in the identification and area extraction of different crops.

Key words: multi-temporal, Sentinel-1A, backscattering coefficient, crop area extraction

Fig. 1

Map of study area and sample distribution"

Table 1

Main parameters of Sentinel-1A SAR data"

获取时间
Acquisition time
入射角
Incidence angle
极化方式
Polarization
产品类型
Product type
分辨率
Resolution (m)
2018/4/17 39.2° VV, VH GRD 5×10
2018/5/05 39.2° VV, VH GRD 5×10
2018/6/16 39.2° VV, VH GRD 5×10
2018/7/22 39.2° VV, VH GRD 5×10
2018/8/27 39.2° VV, VH GRD 5×10
2018/9/02 39.2° VV, VH GRD 5×10

Table 2

Jeffries-Matusita distance of typical crops with different phases"

类型
Type
2018/04/17 2018/05/05 2018/06/16 2018/07/22 2018/08/27 2018/09/02
Sigma Gamma Sigma Gamma Sigma Gamma Sigma Gamma Sigma Gamma Sigma Gamma
棉花-玉米Cotton-corn 0.01 0.01 0.09 0.09 0.53 0.62 0.58 0.59 0.81 0.81 0.12 0.12
棉花-果园Cotton-orchard 0.19 0.18 0.06 0.06 0.29 0.26 0.06 0.06 0.07 0.07 0.40 0.40
棉花-水体Cotton-water 0.97 0.98 1.52 1.51 1.98 1.96 1.76 1.77 1.99 1.99 1.93 1.92
棉花-建筑Cotton-building 1.65 1.65 1.84 1.83 1.70 1.67 1.66 1.65 1.67 1.67 1.64 1.64
玉米-果园Corn-orchard 0.14 0.14 0.04 0.04 1.22 1.22 0.93 0.94 1.15 1.15 0.12 0.12
玉米-水体Corn-water 1.04 1.05 1.49 1.49 1.98 1.94 1.34 1.33 1.99 1.99 1.97 1.97
玉米-建筑Corn-building 1.65 1.65 1.80 1.80 1.89 1.87 1.86 1.86 1.91 1.91 1.50 1.50
果园-水体Orchard-water 1.40 1.40 1.61 1.61 1.99 1.99 1.88 1.88 1.99 1.99 1.99 1.99
果园-建筑Orchard-building 1.55 1.55 1.77 1.77 1.70 1.63 1.66 1.65 1.74 1.74 1.54 1.54
建筑-水体Building-water 1.90 1.90 1.99 1.99 1.99 1.99 1.96 1.96 1.99 1.99 1.98 1.99

Table 3

Jeffries-Matusita distance of typical crops with different band combinations"

类型
Type
6/16 06/16+08/27 06/16+09/02 08/27+09/02 6/16+08/27+09/02
Sigma Gamma Sigma Gamma Sigma Gamma Sigma Gamma Sigma Gamma
棉花-玉米Cotton-corn 0.53 0.62 0.88 0.93 0.83 0.81 1.02 0.93 1.50 1.49
棉花-果园Cotton-orchard 0.29 0.26 0.55 0.54 1.10 1.09 0.85 0.86 1.47 1.45
棉花-水体Cotton-water 1.98 1.96 1.99 1.98 1.98 1.99 1.99 1.97 1.99 1.99
棉花-建筑Cotton-building 1.70 1.67 1.78 1.77 1.77 1.80 1.80 1.78 1.96 1.92
玉米-果园Corn-orchard 1.22 1.22 1.35 1.34 1.35 1.37 1.40 1.38 1.52 1.49
玉米-水体Corn-water 1.98 1.94 1.99 1.96 1.99 1.98 1.98 1.97 1.99 1.98
玉米-建筑Corn-building 1.89 1.87 1.84 1.83 1.78 1.74 1.73 1.70 1.81 1.81
果园-水体Orchard-water 1.99 1.99 1.99 1.98 1.99 1.99 1.98 1.99 1.99 1.99
果园-建筑Orchard-building 1.70 1.63 1.78 1.76 1.65 1.63 1.67 1.66 1.73 1.73
建筑-水体Building-water 1.99 1.99 1.99 1.99 1.99 1.98 1.99 1.99 1.99 1.99

Fig. 2

Technology road map"

Fig. 3

Histograms of backscattering of typical ground objects in multi-temporal SAR data VV: the incident wave is vertically polarized and scatters vertically polarized; VH: the incident wave is vertically polarized and scatters horizontally polarized."

Fig. 4

Time series variation of typical ground objects backscattering features VV: the incident wave is vertically polarized and scatters vertically polarized; VH: the incident wave is vertically polarized and scatters horizontally polarized."

Fig. 5

Classification result of multi-temporal and dual-polarization SAR images ANN: Artificial neural network; SVM: Support vector machine; RF: Random forest."

Table 4

Planting area of different types of crop"

类型
Type
面积
Area (hm2)
占全图比例
Proportion of total graph (%)
棉花Cotton 228842 30.23
玉米Corn 45183 9.47
果园Orchard 71665 5.97

Table 5

Accuracy verification of different classification methods for time series data"

类型
Type
人工神经网络 ANN 支持向量机 SVM 随机森林 RF
Sigma Gamma Sigma Gamma Sigma Gamma
Prod.
Acc
User.
Acc
Prod.
Acc
User.
Acc
Prod.
Acc
User.
Acc
Prod.
Acc
User.
Acc
Prod.
Acc
User.
Acc
Prod.
Acc
User.
Acc
果园Orchard 90.92 73.86 90.78 81.01 89.03 84.55 92.23 84.67 93.17 90.93 91.50 85.42
玉米Corn 17.31 93.66 46.35 88.01 56.45 88.54 48.51 91.65 71.60 93.19 53.38 89.70
棉花Cotton 85.96 86.72 79.06 94.85 89.53 82.11 81.34 99.89 90.88 88.43 86.52 92.61
建筑Building 84.11 98.74 88.57 93.85 84.02 98.74 85.65 90.45 87.14 97.99 82.32 99.57
水体Water 82.07 99.32 44.60 83.42 86.57 97.88 73.52 97.39 92.68 98.01 75.49 99.38
其他Other 97.30 58.18 98.09 52.71 95.74 70.50 98.23 59.06 95.46 85.62 98.79 80.72
Overall. Acc 77.85% 76.42% 84.36% 81.3% 88.97% 82.57%
Kappa 0.73 0.71 0.81 0.77 0.87 0.79

Table 6

Accuracy verification of cotton planting area"

编号
No.
棉花样方面积
Sample area of
cotton (hm2)
棉花分类面积
Classified area of
cotton (hm2)
精度
Accuracy
(%)
编号
No.
非棉花样方面积
Sample area of
non-cotton (hm2)
非棉花分类面积Classified area of non-cotton (hm2) 精度
Accuracy
(%)
1 43.37 43.58 99.53 21 1.40 1.39 98.76
2 102.52 104.98 97.60 22 1.74 1.77 98.35
3 37.12 35.69 96.14 23 2.19 2.26 96.97
4 34.69 33.20 95.69 24 1.24 0.88 95.40
5 31.38 32.88 95.23 25 2.08 1.98 95.14
6 8.44 8.87 94.90 26 32.20 33.82 94.98
7 54.79 51.59 94.15 27 31.13 33.02 93.95
8 8.24 7.51 91.20 28 24.27 21.51 88.62
9 106.06 116.13 90.51 29 16.46 14.44 87.75
10 117.49 129.85 89.48 30 1.71 1.43 83.84
11 131.91 114.47 86.78 31 36.74 29.42 80.08
12 28.15 24.25 86.13 32 72.66 52.31 72.00
13 36.63 31.53 86.08 33 1.18 0.85 71.94
14 105.81 90.80 85.81 34 4.28 5.60 69.09
15 94.85 80.10 84.44 35 21.44 13.70 63.89
16 26.81 21.82 81.40 36 47.88 30.51 63.72
17 18.27 11.80 64.60 37 47.44 26.79 56.47
18 38.16 17.76 46.54 38 3.69 5.42 53.42
19 18.51 8.59 46.40 39 30.36 7.85 25.87
20 28.68 12.81 44.67 40 6.67 1.19 17.86
棉花样方总面积Total sample area (hm2) 1147.85 非棉花样方总面积Total sample area (hm2) 434.81
棉花分类总面积Total classified area (hm2) 1058.03 非棉花分类总面积Total classified area (hm2) 334.89
总精度 Overall accuracy: 92.17% 总精度 Overall accuracy: 73.64%
[1] 杨邦杰, 裴志远. 农作物长势的定义与遥感监测. 农业工程学报, 1999,15(3):214-218.
Yang B J, Pei Z Y. Definition of crop condition and crop monitoring using remote sensing. Trans CSAE, 1999,15(3):214-218 (in Chinese with English abstract).
[2] 陈仲新, 任建强, 唐华俊, 史云, 冷佩, 刘佳, 王利民, 吴文斌, 姚艳敏, 哈斯图亚. 农业遥感研究应用进展与展望. 遥感学报, 2016,20:748-767.
Chen Z X, Ren J Q, Tang H J, Shi Y, Leng P, Liu J, Wang L M, Wu W B, Yao Y M, Hasituya. Progress and perspectives on agricultural remote sensing research and applications in China. J Remote Sens, 2016,20:748-767 (in Chinese with English abstract).
[3] 郭交, 朱琳, 靳标. 基于Sentinel-1和Sentinel-2数据融合的农作物分类. 农业机械学报, 2018,49(4):192-198.
Guo J, Zhu L, Jin B. Crop classification based on data fusion of Sentinel-1 and Sentinel-2. Trans CSAM, 2018,49(4):192-198 (in Chinese with English abstract).
[4] 周涛, 潘剑君, 韩涛, 魏善宝. 基于多时相合成孔径雷达与光学影像的冬小麦种植面积提取. 农业工程学报, 2017,33(10):215-221.
Zhou T, Pan J J, Han T, Wei S B. Planting area extraction of winter wheat based on multi-temporal SAR data and optical imagery. Trans CSAE, 2017,33(10):215-221 (in Chinese with English abstract).
[5] 王利民, 刘佳, 杨福刚, 杨玲波, 姚保民, 王小龙. 基于GF-1卫星遥感数据识别京津冀冬小麦面积. 作物学报, 2018,44:762-773.
Wang L M, Liu J, Yang F G, Yang L B, Yao B M, Wang X L. Acguisition of Winter Wheat Area in the Beijing-Tianjin-Hebei Region with GF-1 Satellite Data. Acta Agron Sin, 2018,44:762-773.
[6] 王迪, 周清波, 陈仲新, 刘佳. 基于合成孔径雷达的农作物识别研究进展. 农业工程学报, 2014,30(16):203-212.
Wang D, Zhou Q B, Chen Z X, Liu J. Research advances on crop identification using synthetic aperture radar. Trans CSAE, 2014,30(16):203-212 (in Chinese with English abstract).
[7] 王松寒, 何隆华. 雷达遥感技术在水稻识别中的研究进展. 遥感信息, 2015,30(2):3-9.
Wang S H, He L H. Advances of rice recognition by SAR. Remote Sens Inf, 2015,30(2):3-9 (in Chinese with English abstract).
[8] 张庆君, 韩晓磊, 刘杰. 星载合成孔径雷达遥感技术进展及发展趋势. 航天器工程, 2017,26(6):1-8.
Zhang Q J, Han X L, Liu J. Technology progress and development trend of space borne synthetic aperture radar remote sensing. Spacecraft Engineer, 2017,26(6):1-8 (in Chinese with English abstract).
[9] Marco P, Alexander B. Assessing fruit-tree crop classification from Landsat-8 time series for the Maipo Valley, Chile. Remote Sens Environ, 2015,171:234-244.
doi: 10.1016/j.rse.2015.10.029
[10] Alexandre B, Thuy L T, Nguyen D L. Monitoring of the rice cropping system in the Mekong Delta using ENVISAT/ASAR dual polarization data. IEEE Trans Geosci Remote Sens, 2009,47:517-526.
doi: 10.1109/TGRS.2008.2007963
[11] Ling F L, Li Z Y, Bai L N, Tian X, Chen E X, Yang Y T. Rice mapping using ALOS PALSAR dual polarization data. J Remote Sens, 2011,15:1215-1227.
[12] Loosvelt L, Peters J, Skriver H, Baets B D, Verhoest N. Impact of reducing polarimetric SAR input on the uncertainty of crop classifications based on the random forests algorithm. IEEE Trans Geosci Remote Sens, 2012,50:4185-4200.
doi: 10.1109/TGRS.2012.2189012
[13] Zhou Y N, Luo J C, Feng L, Yang Y P, Chen Y H, Wu W. Long-short-term-memory-based crop classification using high- resolution optical images and multi-temporal SAR data. GISci Remote Sens, 2019,56:1170-1191.
doi: 10.1080/15481603.2019.1628412
[14] Villa P, Stroppiana D, Fontanelli G, Azar R, Alessandro P B. In-season mapping of crop type with optical and X-band SAR data: a classification tree approach using synoptic seasonal features. Remote Sens, 2015,7:12859-12886.
doi: 10.3390/rs71012859
[15] 李俐, 孔庆玲, 王鹏新, 王蕾, 荀兰. 基于时间序列Sentinel-1A数据的玉米种植面积监测研究. 资源科学, 2018,40:1608-1621.
doi: 10.18402/resci.2018.08.11
Li L, Kong Q L, Wang P X, Wang L, Xun L. Monitoring of maize planting area based on time-series Sentinel-1A SAR data. Resour Sci, 2018,40:1608-1621 (in Chinese with English abstract).
doi: 10.18402/resci.2018.08.11
[16] Jia K, Li Q A, Tian Y C, Wu B F, Zhang F F, Meng J H. Crop classification using multi-configuration SAR data in the North China plain. Int J Remote Sens, 2012,33:170-183.
doi: 10.1080/01431161.2011.587844
[17] 郭交, 尉鹏亮, 周正舒, 苏宝峰. 基于时变特征的多时相PolSAR农作物分类方法. 农业机械学报, 2017,48(12):174-182.
Guo J, Wei P L, Zhou Z S, Su B F. Crop classification method with differential characteristics based on multi-temporal PolSAR Images. Trans CSAM, 2017,48(12):174-182 (in Chinese with English abstract).
[18] 东朝霞, 王迪, 周清波, 陈仲新, 刘佳. 基于SAR遥感的北方旱地秋收作物识别研究. 中国农业资源与区划, 2016,37(8):27-36.
Dong Z X, Wang D, Zhou Q B, Chen Z X, Liu J. Dryland crop identification based on synthetic aperture radar in the North China Plain. Chin J Agric Resour Regional Plan, 2016,37(8):27-36 (in Chinese with English abstract).
[19] 茹皮亚·西拉尔, 杨辽. 新疆棉花物候时空变化遥感监测及气温影响分析. 遥感技术与应用, 2018,33(5):151-159.
Rupiya X, Yang L. Monitoring spatial-temporal change of cotton phenology in Xinjiang and its response to climate change. Remote Sens Technol Appl, 2018,33(5):151-159 (in Chinese with English abstract).
[20] 杜伟娜, 徐爱功, 宋耀鑫, 孙华生. 新型SAR传感器一级地距产品绝对辐射定标方法. 国土资源遥感, 2016,28(4):30-34.
doi: 10.6046/gtzyyg.2016.04.05
Du W N, Xu A G, Song Y X, Sun H S. Absolute radiometric calibration of level-1 detected ground range products of new SAR sensors. Remote Sens Land Resour, 2016,28(4):30-34 (in Chinese with English abstract).
doi: 10.6046/gtzyyg.2016.04.05
[21] Kumar P, Prasad R, Mishra V N, Gupta D K, Singh S K. Artificial neural network for crop classification using C-band RISAT-1 satellite datasets. Russ Agric Sci, 2016,42:281-284.
doi: 10.3103/S1068367416030137
[22] Vapnik V N. Statistical Learning Theory. New York: John Wiley, 1998. pp 421-445.
[23] Chapelle O, Haffner P, Vapnik V N. Support vector machines for histogram-based image classification. IEEE Trans Neural Netw, 1999,10:1055-1064.
doi: 10.1109/72.788646 pmid: 18252608
[24] Leo B. Random forests. Mach Learn, 2001,45:5-32.
doi: 10.1023/A:1010933404324
[25] 化国强. 基于全极化SAR数据玉米长势监测及制图研究. 南京信息工程大学硕士学位论文, 江苏南京, 2011.
Hua G Q. Study on Maize Growth Monitoring and Maize Mapping Based on Full-Polarization SAR data. MS Thesis of Nanjing University of Information Engineering, Nanjing, Jiangsu, China, 2011 (in Chinese with English abstract).
[26] Richards J A, Jia X P. Remote Sensing Digital Image Analysis: An Introduction. Berlin: Springer Verlag, 2006. pp 250-273.
[27] 王凯, 赵军, 朱国锋, 张佩云, 刘江涛, 陈栋栋. 基于GF-1遥感数据决策树与混合像元分解模型的冬小麦种植面积早期估算. 遥感技术与应用, 2018,33(1):158-167.
Wang K, Zhao J, Zhu G F, Zhang P Y, Liu J T, Chen D D. Early estimation of winter wheat planting area in Qingyang city by decision tree and pixel unmixing methods based on GF-1 satellite data. Remote Sens Technol Appl, 2018,33(1):158-167 (in Chinese with English abstract).
[28] 王安琪, 周德民, 宫辉力. 基于雷达后向散射特性进行湿地植被识别与分类的方法研究. 遥感信息, 2012, (2):15-19.
Wang Q A, Zhou D M, Gong H L. A research on method of wetland vegetation identification and classification based on radar backscatter characteristics. Remote Sens Inf, 2012, (2):15-19 (in Chinese with English abstract).
[29] 丁娅萍, 陈仲新. 基于最小距离法的RADARSAT-2遥感数据旱地作物识别. 中国农业资源与区划, 2014,35(6):79-84.
Ding Y P, Chen Z X. Dry land crop classification using minimum distance method based on RADARSAT- 2 data. Chin J Agric Resour Regional Plan, 2014,35(6):79-84 (in Chinese with English abstract).
[30] 郝卫平, 梅旭荣, 蔡学良, 杜建涛, 刘勤. 基于多时相遥感影像的东北三省作物分布信息提取. 农业工程学报, 2011,27(1):201-207.
Hao W P, Mei X R, Cai X L, Du J T, Liu Q. Crop planting extraction based on multi-temporal remote sensing data in Northeast China. Trans CSAE, 2011,27(1):201-207 (in Chinese with English abstract).
[31] Zhang M W, Zhou Q B, Chen Z X, Liu J, Zhou Y, Cai C F. Crop discrimination in Northern China with double cropping systems using Fourier analysis of time-series MODIS data. Int J Appl Earth Observ Geoinf, 2008,10:476-485.
doi: 10.1016/j.jag.2007.11.002
[32] Muhammad S, Zhan Y L, Niu Z, Wang L, Hao P Y. Analyzing the sensitivity of crops classification accuracy based on MODIS EVI time series and history ground reference data. Can J Remote Sens, 2015,41:536-546.
doi: 10.1080/07038992.2015.1112727
[33] 王利民, 杨玲波, 刘佳, 杨福刚, 姚保民. GF-1和MODIS影像冬小麦长势监测指标NDVI的对比. 作物学报, 2018,44:1043-1054.
doi: 10.3724/SP.J.1006.2018.01043
Wang L M, Yang L B, Liu J, Yang F G, Yao B M. Comparison of growth monitoring index NDVI between GF-1 and MODIS images in winter wheat. Acta Agron Sin, 2018,44:1043-1054 (in Chinese with English abstract).
doi: 10.3724/SP.J.1006.2018.01043
[34] 李震, 廖静娟. 合成孔径雷达地表参数反演模型与方法. 北京: 科学出版社, 2011. pp 262-304.
Li Z, Liao J J. Synthetic Aperture Radar Surface Parameters Inversion Model and Methods. Beijing: Science Press, 2011. pp 262-304(in Chinese).
[35] 黎夏, 叶嘉安, 王树功, 刘凯, 刘小平, 钱峻屏, 陈晓越, 何执兼, 覃朝锋. 红树林湿地植被生物量的雷达遥感估算. 遥感学报, 2006,10:387-396.
doi: 10.11834/jrs.20060359
Li X, Ye J A, Wang S G, Liu K, Liu X P, Qian J P, Chen X Y, He Z J, Qin C F. Estimating mangrove wetland biomass using radar remote. sensing. J Remote Sens, 2006,10:387-396 (in Chinese with English abstract).
doi: 10.11834/jrs.20060359
[1] CHEN Bing, WANG Ke-Ru, LI Shao-Hun, XIAO Chun-Hua, SU Yi, TANG Jiang, CHEN Jiang-Lu, JIN Xiu-Liang, LV Yin-Liang, DIAO Mo-Yang, WANG Jie. Monitoring Cotton Field with Suspected Verticillium wilt Using Satellite Remote Sensing with TM Satellite Image as an Example [J]. Acta Agron Sin, 2012, 38(01): 129-139.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!