基于高分三号卫星数据与Η/Α/α-分解特征参数的农作物分类研究

doi:10.3969/j.issn.1004-1524.2022.11.18

浙江农业学报 ›› 2022, Vol. 34 ›› Issue (11): 2491-2503.DOI: 10.3969/j.issn.1004-1524.2022.11.18

基于高分三号卫星数据与Η/Α/α-分解特征参数的农作物分类研究

赵丽仙(), 张王菲(), 李云, 张庭苇, 黄国然

西南林业大学林学院,云南昆明 650224

收稿日期:2021-09-09 出版日期:2022-11-25 发布日期:2022-11-29
通讯作者: 张王菲
作者简介:*张王菲,E-mail:mekmzwf@163.com
赵丽仙(1996—),女,白族,云南大理人,硕士研究生,主要从事资源环境遥感研究。E-mail:zhaolx1108@foxmail.com
基金资助:
国家自然科学基金地区科学基金(42161059);国家自然科学基金地区科学基金(30860240);国家自然科学基金地区科学基金(32160365);高分辨率对地观测系统重大专项(21-Y20B01-9001-19/22-1);云南省万人计划“青年拔尖人才”专项(80201444)

Crops classification based on GF-3 satellite data and Η/Α/α- decomposition characteristic parameters

ZHAO Lixian(), ZHANG Wangfei(), LI Yun, ZHANG Tingwei, HUANG Guoran

College of Forestry, Southwest Forestry University, Kunming 650224, China

Received:2021-09-09 Online:2022-11-25 Published:2022-11-29
Contact: ZHANG Wangfei

摘要/Abstract

摘要：

利用高分三号(GF-3)卫星的全极化C波段多极化合成孔径雷达数据,基于Η/Α/ $α -$ 极化分解提取香农熵(SE)及其强度分量(SE_I)和极化分量(SE_P)、单次反射特征值相对差异度(SERD)、二次反射特征值相对差异度(DERD)、极化比(PF)、基准高度(PH)、极化不对称性(PA)和雷达植被指数(RVI)共9个特征参数,将其应用于农作物分类研究中,以支持向量机(SVM)和随机森林(RF)算法为例,初步探索了基于Η/Α/ $α -$ 分解提取的这9个特征参数在GF-3数据支持下的农作物分类潜力。结果显示:单独将SERD、PH、PF、RVI和SE_P参数用于2种分类方法时,分类精度较高,在82%~92%;但单独运用PA、DERD、SE和SE_I的分类精度均低于80%。将分类精度较低的4个参数组合后,分类精度明显提高,在SVM和RF下的总体分类精度分别达到93.02%和92.05%,Kappa系数均大于0.8。结果表明,基于全极化GF-3数据和Η/Α/ $α -$ 极化分解方法提取的9个特征参数,能很好地表征农作物的散射特征,可用于农作物分类研究。

关键词: 扩展参数, 高分三号卫星, 农作物分类

Abstract:

Based on the civil C-band synthetic aperture radar data of GF-3 satellite, Shannon entropy (SE) and its intensity components (SE_I) and polarization components (SE_P), single bounce eigenvalue relative difference (SERD), double bounce eigenvalue relative difference (DERD), polarization fraction (PF), pedestal height (PH), polarimetric asymmetry (PA) and radar vegetation index (RVI) were extracted from GF-3 full polarimetric data by Η/Α/ $α -$ polarization decomposition method. With application of support vector machine (SVM) and random forest (RF), the potential of these parameters extracted by Η/Α/ $α -$ decomposition from GF-3 data in crops classification was preliminarily explored. The results showed that when SERD, PH, PF, RVI and SE_P parameters were used alone, the classification accuracies were relatively high (82%-92%). When PA, DERD, SE and SE_I parameters were used alone, the classification accuracy was less than 80%. However, the combination of PA, DERD, SE and SE_I parameters improved the classification accuracies to 93.02% and 92.05%, respectively, under SVM and RF, and the Kappa coefficients were all greater than 0.8. Therefore, the nine characteristic parameters extracted by Η/Α/ $α -$ polarization decomposition from GF-3 data could well characterize the scattering characteristics of crops and could be used in crops classification.

Key words: extended parameters, GF-3 satellite, crops classification

中图分类号:

S127
P237

赵丽仙, 张王菲, 李云, 张庭苇, 黄国然. 基于高分三号卫星数据与Η/Α/α-分解特征参数的农作物分类研究[J]. 浙江农业学报, 2022, 34(11): 2491-2503.

ZHAO Lixian, ZHANG Wangfei, LI Yun, ZHANG Tingwei, HUANG Guoran. Crops classification based on GF-3 satellite data and Η/Α/α- decomposition characteristic parameters[J]. Acta Agriculturae Zhejiangensis, 2022, 34(11): 2491-2503.

图/表 12

图1 研究区位置示意图

Fig.1 Schematic diagram of location of study area

图2 基于Η/Α/ α -分解和GF-3极化数据的农作物分类技术路线图

Fig.2 Flowchart of crop classification based on Η/Α/ α - decomposition parameters and GF-3 polarimetric images

图3 基于Η/Α/ α -分解的9个参数的极化散射特征统计图 DN,遥感影像像元亮度值。

Fig.3 Statistical diagram of polarization scattering characteristics of 9 parameters derived fromΗ/Α/ α - decomposition DN, Digital number, the photopixel brightness value of the remote sensing image.

表1 极化散射特征统计量

Table 1 List of polarization scattering characteristic statistics

参数 Parameter	小麦 Wheat	油菜 Rape	植被 Vegetation
DERD	0.62±0.06	0.55±0.03	0.57±0.03
PF	0.69±0.06	0.32±0.06	0.49±0.05
PA	0.78±0.06	0.56±0.14	0.71±0.08
PH	0.14±0.03	0.48±0.07	0.29±0.04
RVI	0.41±0.08	0.91±0.09	0.69±0.07
SERD	0.72±0.07	0.32±0.08	0.53±0.05
SE	1.54±0.43	0.58±0.32	2.92±0.96
SE_I	0.53±0.41	0.67±0.36	2.49±0.94
SE_P	1.06±0.22	0.15±0.06	0.43±0.11

图4 分割前后的各参数影像

Fig.4 Images of various parameters before and after segmentation

图5 基于单一参数的SVM分类结果

Fig.5 SVM classification results based on single parameter

图6 单一参数的RF分类结果

Fig.6 RF classification results based on single parameter

图7 基于单一参数的SVM和RF分类的总体精度

Fig.7 Overall accuracy of SVM and RF classification based on single parameter

表2 基于单一参数的SVM和RF分类的Kappa系数

Table 2 Kappa coefficient of SVM and RF classification based on single parameter

参数Parameter	SVM	RF
DERD	0.33	0.23
PF	0.74	0.81
PA	0.43	0.34
PH	0.81	0.86
RVI	0.86	0.82
SERD	0.78	0.84
SE	0.58	0.57
SE_I	0.40	0.32
SE_P	0.89	0.87

图8 基于不同参数进行SVM和RF分类的生产者精度和用户精度

Fig.8 Producer’s accuracy and user’s accuracy of SVM and RF classification based on single parameter

表3 组合参数的分类精度

Table 3 Classification accuracy of combined parameters

参数组合 Combination of parameters	总体精度 Overall accuracy/%		Kappa系数 Kappa coefficient
参数组合 Combination of parameters	SVM	RF	SVM	RF
PA+SE	78.80	74.03	0.68	0.60
PA+SE_I	82.90	82.82	0.74	0.74
PA+SE+SE_I	86.00	86.44	0.79	0.80
PA+DERD+SE+SE_I	93.02	92.05	0.89	0.88

图9 基于参数组合的分类结果

Fig.9 SVM and RF classification result based on combined parameters

参考文献 36

[1]	高晗, 汪长城, 杨敏华, 等. 基于高分三号极化SAR数据的农作物散射特性分析及分类[J]. 测绘工程, 2019, 28(3): 50-56.
	GAO H, WANG C C, YANG M H, et al. Scattering feature analysis and classification of crops based on GF-3 PolSAR data[J]. Engineering of Surveying and Mapping, 2019, 28(3): 50-56. (in Chinese with English abstract)
[2]	杨劲松, 任林, 王隽. 高分三号卫星对海浪的首次定量遥感[J]. 海洋与湖沼, 2017, 48(2): 207-209.
	YANG J S, REN L, WANG J. The first quantitative remote sensing of ocean surface waves by Chinese GF-3 SAR satellite[J]. Oceanologia et Limnologia Sinica, 2017, 48(2): 207-209. (in Chinese with English abstract)
[3]	王文煜, 谢春华, 袁新哲, 等. 高分三号卫星ATI模式海表面流场测量性能分析[J]. 航天器工程, 2017, 26(1): 132-139.
	WANG W Y, XIE C H, YUAN X Z, et al. Performance analysis of sea surface current measurements by GF-3 satellite ATI mode[J]. Spacecraft Engineering, 2017, 26(1): 132-139. (in Chinese with English abstract)
[4]	刘泽宇, 柳彬, 郭炜炜, 等. 高分三号NSC模式SAR图像舰船目标检测初探[J]. 雷达学报, 2017, 6(5): 473-482.
	LIU Z Y, LIU B, GUO W W, et al. Ship detection in GF-3 NSC mode SAR images[J]. Journal of Radars, 2017, 6(5): 473-482. (in Chinese with English abstract)
[5]	李平湘, 赵伶俐, 任烨仙. 合成孔径雷达在农业监测中的应用和展望[J]. 地理空间信息, 2017, 15(3): 1-4.
	LI P X, ZHAO L L, REN Y X. Outlook and application of the synthetic aperture radar in agriculture monitoring[J]. Geospatial Information, 2017, 15(3): 1-4. (in Chinese with English abstract)
[6]	秦昆, 陈一祥, 甘顺子, 等. 高分辨率遥感影像空间结构特征建模方法综述[J]. 中国图象图形学报, 2013, 18(9): 1055-1064.
	QIN K, CHEN Y X, GAN S Z, et al. Review on methods of spatial structural feature modeling of high resolution remote sensing images[J]. Journal of Image and Graphics, 2013, 18(9): 1055-1064. (in Chinese with English abstract)
[7]	韩震, 金亚秋. ERS-2 SAR和Landsat ETM+数据融合提取崇明东滩典型地物信息与分类[J]. 海洋环境科学, 2006, 25(3): 21-24.
	HAN Z, JIN Y Q. Information extraction and classification of typical canopies over the Chongming eastern tidal flat from data fusion of ERS-2 SAR and Landsat ETM+[J]. Marine Environmental Science, 2006, 25(3): 21-24. (in Chinese with English abstract)
[8]	曾妍, 王迪, 田甜, 等. 极化SAR农作物分类研究进展[J]. 中国农业信息, 2020, 32(2): 13-26.
	ZENG Y, WANG D, TIAN T, et al. Research advances on crop classification using PolSAR data[J]. China Agricultural Informatics, 2020, 32(2): 13-26. (in Chinese with English abstract)
[9]	韩宇. 多时相C波段全极化SAR农作物识别方法研究[D]. 呼和浩特: 内蒙古师范大学, 2015.
	HAN Y. Research of crop recognition method based on multi-temporal C-band quad-poi SAR[D]. Hohhot: Inner Mongolia Normal University, 2015. (in Chinese with English abstract)
[10]	邹斌, 张腊梅, 孙德明, 等. PolSAR图像信息提取技术及应用的发展[J]. 遥感技术与应用, 2009, 24(3): 263-273.
	ZOU B, ZHANG L M, SUN D M, et al. Development and application of information extraction using polarimetric SAR data[J]. Remote Sensing Technology and Application, 2009, 24(3): 263-273. (in Chinese with English abstract)
[11]	张腊梅, 段宝龙, 邹斌. 极化SAR图像目标分解方法的研究进展[J]. 电子与信息学报, 2016, 38(12): 3289-3297.
	ZHANG L M, DUAN B L, ZOU B. Research development on target decomposition method of polarimetric SAR image[J]. Journal of Electronics & Information Technology, 2016, 38(12): 3289-3297. (in Chinese with English abstract)
[12]	LEE J S, POTTIER E. 极化雷达成像基础与应用[M].洪文, 李洋, 尹嫱, 等,译. 北京: 电子工业出版社, 2013.
[13]	JIAO X F, KOVACS J M, SHANG J L, et al. Object-oriented crop mapping and monitoring using multi-temporal polarimetric RADARSAT-2 data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2014, 96: 38-46. DOI URL
[14]	LI H P, ZHANG C, ZHANG S Q, et al. Full year crop monitoring and separability assessment with fully-polarimetric L-band UAVSAR: a case study in the Sacramento Valley, California[J]. International Journal of Applied Earth Observation and Geoinformation, 2019, 74: 45-56. DOI URL
[15]	张继超, 郭伟, 周沛希. 雷达植被指数与香农熵在GF-3影像分类中的应用[J]. 测绘科学, 2020, 45(10): 48-54.
	ZHANG J C, GUO W, ZHOU P X. Application of radar vegetation index and Shannon entropy in GF-3 image classification[J]. Science of Surveying and Mapping, 2020, 45(10): 48-54. (in Chinese with English abstract)
[16]	马腾, 王耀强, 李瑞平, 等. 基于微波遥感极化目标分解的土地覆盖/土地利用分类[J]. 农业工程学报, 2015, 31(2): 259-265.
	MA T, WANG Y Q, LI R P, et al. Land cover/land use classification based on polarimetric target decomposition of microwave remote sensing[J]. Transactions of the Chinese Society of Agricultural Engineering, 2015, 31(2): 259-265. (in Chinese with English abstract)
[17]	姚天宇. 长征四号丙运载火箭成功发射高分三号卫星[J]. 中国航天, 2016(8): 8.
	YAO T Y. The CZ-4 carrier rocket successfully launched GF-3 satellite[J]. Aerospace China, 2016(8): 8. (in Chinese)
[18]	JI Y J, XU K P, ZENG P, et al. GA-SVR algorithm for improving forest above ground biomass estimation using SAR data[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 6585-6595. DOI URL
[19]	张王菲, 姬永杰. 极化与干涉SAR植被参数反演[M]. 北京: 中国林业出版社, 2019.
[20]	ALLAIN S, FERRO-FAMIL L, POTTIER E. Two novel surface model based inversion algorithms using multi-frequency polSAR data[C]// IGARSS 2004. 2004 IEEE International Geoscience and Remote Sensing Symposium. Anchorage, AK, USA: IEEE, 2004: 823-826.
[21]	ALLAIN S, FERROFAMIL L, POTIER E. New eigenvalue-based parameters for natural media characterization[C]// European Radar Conference, 2005. EURAD 2005. Paris, France: IEEE, 2005: 177-180.
[22]	RÉFRÉGIER P, MORIO J. Shannon entropy of partially polarized and partially coherent light with Gaussian fluctuations[J]. Journal of the Optical Society of America A, Optics, Image Science, and Vision, 2006, 23(12): 3036-3044. DOI URL
[23]	AINSWORTH T L, LEE J S, SCHULER D L. Multi-frequency polarimetric SAR data analysis of ocean surface features[C]// IGARSS 2000. IEEE 2000 International Geoscience and Remote Sensing Symposium. Honolulu, HI, USA: IEEE, 2000: 1113-1115.
[24]	AINSWORTH T L, CLOUDE S R, LEE J S. Eigenvector analysis of polarimetric SAR data[C]// IEEE International Geoscience and Remote Sensing Symposium. Toronto, ON, Canada: IEEE, 2002: 626-628.
[25]	VAN ZYL J J. Application of Cloude’s target decomposition theorem to polarimetric imaging radar data[J]. Proceedings of SPIE, 1993, 1748: 184-191.
[26]	DURDEN S L, VAN ZYL J J, ZEBKER H A. The unpolarized component in polarimetric radar observations of forested areas[J]. IEEE Transactions on Geoscience and Remote Sensing, 1990, 28(2): 268-271. DOI URL
[27]	WANG P Z. Pattern recognition with fuzzy objective function algorithms[J]. SIAM Review, 1983, 25(3): 442.
[28]	王馨爽. 多维度SAR森林及其类型分类识别方法研究[D]. 西安: 西安科技大学, 2013.
	WANG X S. Multi-dimension SAR for forest recognition and forest type classification[D]. Xi’an: Xi’an University of Science and Technology, 2013. (in Chinese with English abstract)
[29]	隋艳立. 基于深度支持向量机的极化SAR图像分类[D]. 西安: 西安电子科技大学, 2014.
	SUI Y L. Polarmetric SAR image classification based on deep SVM[D]. Xi’an: Xidian University, 2014. (in Chinese with English abstract)
[30]	HE F C, GU L J, REN R Z, et al. Research on crop classification in northeast China based on multi-source and multi-temporal SAR Images[J]. Proceedings of SPIE, 2018, 10767: 124-136.
[31]	CHAN J C W, PAELINCKX D. Evaluation of Random Forest and Adaboost tree-based ensemble classification and spectral band selection for ecotope mapping using airborne hyperspectral imagery[J]. Remote Sensing of Environment, 2008, 112(6): 2999-3011. DOI URL
[32]	LIU C, SHANG J L, VACHON P W, et al. Multiyear crop monitoring using polarimetric RADARSAT-2 data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2013, 51(4): 2227-2240. DOI URL
[33]	朱凤敏, 吴迪, 杨佳琪. 基于Sentinel-1B SAR数据的农作物分类方法研究[J]. 测绘与空间地理信息, 2020, 43(5): 105-108.
	ZHU F M, WU D, YANG J Q. Research on crop classification method based on sentinel-1B SAR data[J]. Geomatics & Spatial Information Technology, 2020, 43(5): 105-108. (in Chinese with English abstract)
[34]	LARRAÑAGA A, ÁLVAREZ-MOZOS J. On the added value of quad-pol data in a multi-temporal crop classification framework based on RADARSAT-2 imagery[J]. Remote Sensing, 2016, 8(4): 335. DOI URL
[35]	TAMIMINIA H, HOMAYOUNI S, MCNAIRN H, et al. A particle swarm optimized kernel-based clustering method for crop mapping from multi-temporal polarimetric L-band SAR observations[J]. International Journal of Applied Earth Observation and Geoinformation, 2017, 58: 201-212. DOI URL
[36]	王晨丞, 王永前, 王利花. 基于SAR纹理信息的农作物识别研究: 以农安县为例[J]. 遥感技术与应用, 2021, 36(2): 372-380.
	WANG C C, WANG Y Q, WANG L H. Crop identification based on SAR texture information: a case study of Nong’an County[J]. Remote Sensing Technology and Application, 2021, 36(2): 372-380. (in Chinese with English abstract)

基于高分三号卫星数据与Η/Α/α-分解特征参数的农作物分类研究

Crops classification based on GF-3 satellite data and Η/Α/α- decomposition characteristic parameters

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 36

相关文章 14

编辑推荐

Metrics

本文评价

[1]	李永梅, 王浩, 赵勇, 张立根. 基于连续统去除法的枸杞叶片含水率高光谱估算[J]. 浙江农业学报, 2022, 34(4): 781-789.
[2]	张矞勋, 王磊, 璩向宁, 曹媛, 吴梦瑶, 于瑞鑫, 孙源. GF-1/WFV在玉米叶面积指数估算中的应用研究[J]. 浙江农业学报, 2021, 33(5): 861-872.
[3]	卢牧原, 刘源, 刘桂建. 基于组合模型的黑土区土壤有机质含量预测分析[J]. 浙江农业学报, 2020, 32(8): 1427-1436.
[4]	钱家炜, 刘晓青, 张静静, 周卫红, 李建龙. 张家港市农田土壤重金属含量高光谱遥感监测模型构建[J]. 浙江农业学报, 2020, 32(8): 1437-1445.
[5]	禹文杰, 王彩霞, 乔芦, 王松磊, 贺晓光. 基于高光谱成像技术的泾源黄牛肉色度PLSR预测模型构建[J]. 浙江农业学报, 2020, 32(3): 527-533.
[6]	杨红云, 周琼, 杨珺, 孙玉婷, 路艳, 殷华. 基于高光谱的水稻叶片氮素营养诊断研究[J]. 浙江农业学报, 2019, 31(10): 1575-1582.
[7]	孟庆龙, 张艳, 尚静. 基于高光谱成像的猕猴桃表面疤痕无损识别[J]. 浙江农业学报, 2019, 31(8): 1372-1378.
[8]	王美玲, 焦琳琳, 王晓红, 吴兵, 肖星星. 曹妃甸湿地典型植被光谱特征差异性分析[J]. 浙江农业学报, 2019, 31(6): 963-969.
[9]	梁长江, 吴雪梅, 王芳, 宋朱军, 张富贵. 基于无人机的田间地膜识别算法研究[J]. 浙江农业学报, 2019, 31(6): 1005-1011.
[10]	金秀, 卢杰, 傅运之, 王帅, 许高健, 李绍稳. 基于深度卷积神经网络的小麦赤霉病高光谱病症点分类方法[J]. 浙江农业学报, 2019, 31(2): 315-325.
[11]	王文才, 赵刘, 李绍稳, 齐海军, 金秀, 王帅. 基于特征波长选择和建模的高光谱土壤总氮含量估测方法研究[J]. 浙江农业学报, 2018, 30(9): 1576-1584.
[12]	余昌乐, 许童羽, 王洋, 于丰华. 双子叶植物叶片类胡萝卜素含量高光谱反演估算[J]. 浙江农业学报, 2018, 30(3): 393-398.
[13]	杨福芹, 冯海宽, 李振海, 杨贵军, 戴华阳. 基于高光谱的苹果叶片叶绿素含量估算[J]. 浙江农业学报, 2017, 29(10): 1742-1748.
[14]	许童羽, 洪雪, 陈春玲, 周云成, 曹英丽, 于丰华, 李娜. 基于冠层NDVI数据的北方粳稻产量模型研究[J]. 浙江农业学报, 2016, 28(10): 1790-1795.