RapidEye卫星红边波段对主要农作物识别能力的影响研究
|
摘要
德国RapidEye卫星在传统的可见光波段和近红外波段的基础上加入了红边波段,红边波段更有利于地表植被分类和监测植被生长状态。为了定量分析RapidEye卫星影像加入红边波段后对地物识别能力的影响,文章利用内蒙古呼伦贝尔市莫力达瓦达斡尔族自治旗哈达阳镇2017年7月25日的RapidEye数据,采用BP神经网络的监督分类方法,提取了研究区内玉米和大豆面积,并计算了有红边和无红边条件下该地区玉米、大豆以及其他作物之间的可分性测度和混淆矩阵,对比了这两种波段组合形式下农作物提取的可分性和基于混淆矩阵的分类精度。结果证明:引入红边波段后,玉米-大豆、大豆-其他、玉米-其他的可分性测度分别从1.71、1.97、1.91提高到1.99、1.99、1.93。总体分类精度从79.57%提高到84.14%,Kappa系数从0.63提高到了0.71。该结果证明,红边波段对区分玉米-大豆的能力有显著提高,区分其他地物的能力也有了明显提升。国外搭载红边波段的卫星载荷越来越多,国产卫星也拟引入红边波段载荷技术,为农业部门提供更可靠的数据支持,文章的研究结果能为国产红边波段数据在农业上的应用提供参考。
On the basis of the traditional visible band and near infrared band, the German Rapid Eye satellite sensor provides red-edge waveband which is more conducive to the classification of surface vegetation and the monitor of vegetation growth. In order to quantitatively analyze the impact of recognition ability which Rapid Eye satellite images adding the red edge band, the area of corn and soybean in the study area was extracted with the monitoring classification method based on BP neural network by using the Rapid Eye satellite remote sensing data of Hadayang town, Hulunbuir, Inner Mongolia on July 25, 2017. Separable measure and confusion matrix of corn, soybean and other kinds of ground objects in the region were calculated under 2 kinds of waveband combinations with or without the involvement of red-edge waveband.Furthermore, the separability of crop extraction and classification accuracy based on confusion matrix was compared and the classification accuracy verification was evaluated by the study area visual interpretation results. The results show that : The degree of separation between corns and soybeans, soybeans and others, as well as corn and others were improved from 1.71 to 1.99, from 1.97 to 1.99, from 1.91 to 1.93, respectively. The overall accuracy was increased from 79.57% to 84.14%,and the Kappa coefficient was increased from 0.63 to 0.71. The combined data obtained in this study demonstrated that the red-edge waveband significantly improved the ability to distinguish corn from soybean and other objects. There are more and more satellite loads on the red-edge waveband in foreign countries. The domestic satellite also plans to introduce the rededge waveband load technology, which provides more reliable data support for the agricultural administration sector. The results of this study provide reference for the application of red-edge waveband data in agriculture.
On the basis of the traditional visible band and near infrared band, the German Rapid Eye satellite sensor provides red-edge waveband which is more conducive to the classification of surface vegetation and the monitor of vegetation growth. In order to quantitatively analyze the impact of recognition ability which Rapid Eye satellite images adding the red edge band, the area of corn and soybean in the study area was extracted with the monitoring classification method based on BP neural network by using the Rapid Eye satellite remote sensing data of Hadayang town, Hulunbuir, Inner Mongolia on July 25, 2017. Separable measure and confusion matrix of corn, soybean and other kinds of ground objects in the region were calculated under 2 kinds of waveband combinations with or without the involvement of red-edge waveband.Furthermore, the separability of crop extraction and classification accuracy based on confusion matrix was compared and the classification accuracy verification was evaluated by the study area visual interpretation results. The results show that : The degree of separation between corns and soybeans, soybeans and others, as well as corn and others were improved from 1.71 to 1.99, from 1.97 to 1.99, from 1.91 to 1.93, respectively. The overall accuracy was increased from 79.57% to 84.14%,and the Kappa coefficient was increased from 0.63 to 0.71. The combined data obtained in this study demonstrated that the red-edge waveband significantly improved the ability to distinguish corn from soybean and other objects. There are more and more satellite loads on the red-edge waveband in foreign countries. The domestic satellite also plans to introduce the rededge waveband load technology, which provides more reliable data support for the agricultural administration sector. The results of this study provide reference for the application of red-edge waveband data in agriculture.
引文
[1]姚云军,秦其明,张自力,等.高光谱技术在农业遥感中的应用研究进展[J].农业工程学报,2008(7):301-306.
[2]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.
[3]王利民,刘佳,杨福刚,等.基于GF-1卫星遥感的冬小麦面积早期识别[J].农业工程学报,2015,31(11):194-201.
[4]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.
[5]黄敬峰,王渊,王福民,等.油菜红边特征及其叶面积指数的高光谱估算模型[J].农业工程学报,2006(8):22-26.
[6]凌春丽,朱兰艳,吴俐民.World View-2影像林地信息提取的研究与实现[J].测绘科学,2010,35(5):205-207.
[7]刘佳,王利民,滕飞,等.Rapid Eye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.
[8]DUPUY S,BARDE E,BALESTRAT M.An object-based image analysis method for monitoring land conversion by artificial sprawl use of Rapid Eye and IRS data[J].Remote Sensing,2012,4(2):404-423.
[9]张振兴,李宁,刘阳.基于Worldview-2多光谱遥感数据纹理特征提取方法[J].系统工程与电子技术,2013,35(10):2044-2049.
[10]RICHTER K,ATZBERGER C,VUOLO F,et al.Evaluation of Sentinel-2 spectral sampling for radiative transfer model based LAI estimation of wheat,sugar beet,and maize[J].IEEE Journal of Selected Topics in Applied Earth Observations&Remote Sensing,2011,4(2):458-464.
[11]KIM H O,YEOM J M.Effect of red-edge and texture features for object-based paddy rice crop classification using Rapid Eye multi-spectral satellite image data[J].International Journal of Remote Sensing,2014,35(19):7046-7068.
[12]KIM H O,YEOM J M.Sensitivity of vegetation indices to spatial degradation of Rapid Eye imagery for paddy rice detection:A case study of South Korea[J].Giscience&Remote Sensing,2015,52(1):1-17.
[13]佘宝,黄敬峰,石晶晶,等.基于红边位置变化特征的油菜种植区域提取[J].农业工程学报,2013,29(15):145-152.
[14]ADELABU S,MUTANGA O,ADAM E.Evaluating the impact of red-edge band from Rapideye image for classifying insect defoliation levels[J].Journal of Photogrammetry&Remote Sensing,2014,95(3):34-41.
[15]RUMELHART,MCCLELL,PARALLEL.distributed proces sing(vol.1)[M].The MIT Press,1986:45-76.
[16]谭玉敏,夏玮.基于最佳波段组合的高光谱遥感影像分类[J].测绘与空间地理信息,2014,37(4):19-22.
[2]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.
[3]王利民,刘佳,杨福刚,等.基于GF-1卫星遥感的冬小麦面积早期识别[J].农业工程学报,2015,31(11):194-201.
[4]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.
[5]黄敬峰,王渊,王福民,等.油菜红边特征及其叶面积指数的高光谱估算模型[J].农业工程学报,2006(8):22-26.
[6]凌春丽,朱兰艳,吴俐民.World View-2影像林地信息提取的研究与实现[J].测绘科学,2010,35(5):205-207.
[7]刘佳,王利民,滕飞,等.Rapid Eye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.
[8]DUPUY S,BARDE E,BALESTRAT M.An object-based image analysis method for monitoring land conversion by artificial sprawl use of Rapid Eye and IRS data[J].Remote Sensing,2012,4(2):404-423.
[9]张振兴,李宁,刘阳.基于Worldview-2多光谱遥感数据纹理特征提取方法[J].系统工程与电子技术,2013,35(10):2044-2049.
[10]RICHTER K,ATZBERGER C,VUOLO F,et al.Evaluation of Sentinel-2 spectral sampling for radiative transfer model based LAI estimation of wheat,sugar beet,and maize[J].IEEE Journal of Selected Topics in Applied Earth Observations&Remote Sensing,2011,4(2):458-464.
[11]KIM H O,YEOM J M.Effect of red-edge and texture features for object-based paddy rice crop classification using Rapid Eye multi-spectral satellite image data[J].International Journal of Remote Sensing,2014,35(19):7046-7068.
[12]KIM H O,YEOM J M.Sensitivity of vegetation indices to spatial degradation of Rapid Eye imagery for paddy rice detection:A case study of South Korea[J].Giscience&Remote Sensing,2015,52(1):1-17.
[13]佘宝,黄敬峰,石晶晶,等.基于红边位置变化特征的油菜种植区域提取[J].农业工程学报,2013,29(15):145-152.
[14]ADELABU S,MUTANGA O,ADAM E.Evaluating the impact of red-edge band from Rapideye image for classifying insect defoliation levels[J].Journal of Photogrammetry&Remote Sensing,2014,95(3):34-41.
[15]RUMELHART,MCCLELL,PARALLEL.distributed proces sing(vol.1)[M].The MIT Press,1986:45-76.
[16]谭玉敏,夏玮.基于最佳波段组合的高光谱遥感影像分类[J].测绘与空间地理信息,2014,37(4):19-22.