基于GF-1/WFV时间序列数据的河套灌区主要农作物识别

doi:10.3969/j.issn.2096-1197.2019.05.21

北方农业学报 ›› 2019, Vol. 47 ›› Issue (5): 112-118.doi: 10.3969/j.issn.2096-1197.2019.05.21

• 农业经济·农业信息技术 • 上一篇下一篇

基于GF-1/WFV时间序列数据的河套灌区主要农作物识别

乌云德吉^1,2, 于利峰¹, 承昊¹, 包珺玮¹, 许洪滔¹, 赵佳乐¹, 乌兰吐雅¹

1.内蒙古自治区农牧业科学院农牧业经济与信息研究所,内蒙古呼和浩特 010031;
2.中国农业科学院农业资源与农业区划研究所/农业农村部农业遥感重点实验室,北京 100081

收稿日期:2019-09-07 出版日期:2019-10-20 发布日期:2019-12-11
作者简介:乌云德吉（1987—）,女,助理研究员,博士研究生,研究方向为农情遥感监测。
基金资助:
内蒙古农牧业青年创新基金项目（2017QNJJN10,2018CXJJ09）

Research on the identification method of main crops in Hetao Irrigated Area based on GF-1/WFV NDVI Time Series Data

Wuyundeji^1,2, YU Lifeng¹, CHENG Hao¹, BAO Junwei¹, XU Hongtao¹, ZHAO Jiale¹, Wulantuya¹

1.Inner Mongolia Academy of Agricultural & Animal Husbandry Sciences,Hohhot 010031,China;;
2.Key Laboratory of Agriinformation,Ministry of Agriculture and Rural Affairs/Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,Beijing 100081,China

Received:2019-09-07 Online:2019-10-20 Published:2019-12-11

摘要/Abstract

摘要： 河套灌区农业资源条件较好,适宜于多种农作物的生长,且地块较为破碎,为精准进行该地区农业资源的调查和农情评价,采用遥感技术精准识别农作物种类。文章利用2017年4—10 月内9个时相的GF-1/WFV 数据,结合实地采集样本点转化成感兴区的方法,通过计算分析春小麦、玉米、向日葵、西葫芦和番茄5种农作物在整个生长阶段的时间序列NDVI值和变化特点,构建了基于NDVI阈值分割的决策树分类模型,对研究区春小麦、玉米、向日葵、西葫芦和番茄5种农作物进行了识别分类;在5 m分辨率的RapidEye 数据上选定了10个验证样方,目视解译样方内作物,利用解译结果对决策树分类结果进行验证,并用混淆矩阵表达。结果表明：春小麦、玉米、向日葵、西葫芦和番茄的用户精度分别为：88.86%,62.44%,87.29 %,65.78 %,总体分类精度达到76.29%,Kappa系数为0.652 9 。结论为通过分析基于GF-1/WFV NDVI时间序列数据可以较为准确识别中尺度上的大宗农作物,该方法适宜于农业资源调查遥感应用和研究。

关键词: 农作物识别, GF-1/WFV, NDVI, 决策树, 光谱信息

Abstract: Agricultural resources are in good condition in Hetao Irrigated Area,suitable for the growth of a variety of crops,and the plots are relatively fragmented.A remote sensing technology can be used to accurately identify crop types so that agricultural resources in this area can be investigated and the agricultural conditions can be evaluated.Using 9 phases of GF-1/WFV data in the period from April 2017 to October 2017,this paper combined the method of transforming the field sample points into sense areas to calculate and analyze the time series NDVI values and variation characteristics of spring wheat,corn,helianthus and vegetable in the whole growth stage in the study area,and thenthe decision tree classification model based on threshold segmentation was constructed to identify and classify the above five crops in this area.Ten verification quadrangles were selected from RapidEye data with a resolution of 5 m,the crops in the quadrangle were visually interpreted,and the classification results of the decision tree were verified with the interpretation results and expressed with confusion matrix.The results showed that identification accuracies of spring wheat,corn,sunflowers,zucchini,and tomatoes were 88.86%,62.44%,87.29% and 65.78%,respectively,while the overall accuracy reached 76.29% and the Kappa coefficient was 0.652 9.It can be seen that by analyzing the time series data based on GF-1/WFV NDVI,large crops in mesoscale can be identified more accurately.This method is suitable for the application and research of remote sensing in agricultural resource survey.

Key words: Crop identification, GF-1/WFV, NDVI, Decision tree, Spectral information

中图分类号:

TP701

乌云德吉, 于利峰, 承昊, 包珺玮, 许洪滔, 赵佳乐, 乌兰吐雅. 基于GF-1/WFV时间序列数据的河套灌区主要农作物识别[J]. 北方农业学报, 2019, 47(5): 112-118.

Wuyundeji, YU Lifeng, CHENG Hao, BAO Junwei, XU Hongtao, ZHAO Jiale, Wulantuya. Research on the identification method of main crops in Hetao Irrigated Area based on GF-1/WFV NDVI Time Series Data[J]. Journal of Northern Agriculture, 2019, 47(5): 112-118.

参考文献

[1] 陈仲新,任建强,唐华俊,等. 农业遥感研究应用进展与展望[J]. 遥感学报,2016,20(5):748-767.
[2] 杨闫君,占玉林,田庆久,等. 基于GF-1/WFV NDVI时间序列数据的作物分类[J]. 农业工程学报,2015,31(24):155-161.
[3] 刘佳,王利民,姚保民,等. 基于多时相OLI数据的宁夏大尺度水稻面积遥感估算[J]. 农业工程学报,2017,33(15):200-209.
[4] 王利民,刘佳,姚保民,等. 基于GF-1影像NDVI年度间相关分析的冬小麦面积变化监测[J]. 农业工程学报,2018,34(8):184-191.
[5] 刘吉凯,钟仕全,梁文海. 基于多时相Landsat8 OLI影像的作物种植结构提取[J]. 遥感技术与应用,2015,30(4):775-783.
[6] 顾晓鹤,宋国宝,韩立建,等. 基于变化向量分析的冬小麦长势变化监测研究[J]. 农业工程学报,2008,24(4):159-165.
[7] 林文鹏,王长耀,储德平,等. 基于光谱特征分析的主要秋季作物类型提取研究[J]. 农业工程学报,2006(9):128-132.
[8] JAKUBAUSKAS M E,LEGATES D R,KASTENS J H.Crop identification using harmonic analysis of time-series AVHRR NDVI data[J].Computers and Electronics in Agriculture,2002,37(1):127-139.
[9] 刘佳,王利民,杨福刚,等. 基于HJ时间序列数据的农作物种植面积估算[J]. 农业工程学报,2015,31(3):199-206.
[10] SOLBERG A H S,TAXT T,JAIN A K. A Markov random field model for classification of multisource satellite imagery[J]. IEEE Transactions on Geoscience and Remote Sensing,1996,34(1):100-113.
[11] TURNER M D,CONGALTON R G.Classification of multi-temporal spot-xs satellite data for mapping rice fields on a west african floodplain[J]. International Journal of Remote Sensing,1998,19(1):21-41.
[12] 陈怀亮,李颖,张红卫. 农作物长势遥感监测业务化应用与研究进展[J]. 气象与环境科学,2015,38(1):95-102.
[13] 侯智惠,梅连杰,侯安宏,等. 内蒙古农业资源利用区域差异分析[J]. 中国农业资源与区划,2016,37(1):160-166.
[14] 刘佳,王利民,腾飞. RapidEye卫星红边波段对农作物面积提取精度的影响[J]. 农业工程学报,2016,32(13):140-148.
[15] JANSSEN L L F,VAN DER WEL F J M.Accuracy assessment of satellite derived land-cover data:a review[J]. Photogrammetric Engineering & Remote Sensing,1994,60(4):419-426.
[16] KALKHAN M A,REICH R M,STOHLGREN T J.Technical note assessing the accuracy of landsat thematic mapper classification using double sampling[J]. International Journal of Remote Sensing,1998,19(11):2049-2060.
[17] 刘晓娜,封志明,姜鲁光. 基于决策树分类的橡胶林地遥感识别[J]. 农业工程学报,2013,29(24):163-172.

基于GF-1/WFV时间序列数据的河套灌区主要农作物识别

Research on the identification method of main crops in Hetao Irrigated Area based on GF-1/WFV NDVI Time Series Data

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 6

Metrics

本文评价

推荐阅读 0

[1]	包珺玮, 于利峰, 乌兰吐雅, 许洪滔, 于伟卓, 敦惠霞. 基于ESTARFM模型的农作物类型识别方法应用[J]. 北方农业学报, 2021, 49(3): 128-134.
[2]	乌云德吉;乌兰吐雅;于利峰;许洪滔;包珺玮. 基于GF-1 WFV数据的河套灌区春小麦长势遥感分析——以临河区为例[J]. 北方农业学报, 2018, 46(2): 123-123.
[3]	乌云德吉;于利峰;包珺玮;许洪滔;乌兰吐雅. RapidEye卫星红边波段对主要农作物识别能力的影响研究[J]. 北方农业学报, 2017, 45(6): 118-118.
[4]	李思慧. 生长季气象条件对通辽地区植被长势影响分析[J]. 北方农业学报, 2017, 45(3): 90-90.
[5]	郭晓丽. 乌兰察布市近十年植被指数时空特征分析[J]. 北方农业学报, 2017, 45(3): 86-86.
[6]	于利峰;乌兰吐雅;乌兰;包珺玮. 基于SAR-C的大兴安岭西麓地区主要农作物识别方法研究[J]. 北方农业学报, 2017, 45(3): 108-108.