基于SPOT-5影像的马尾松毛虫虫害遥感监测研究
|
摘要
马尾松毛虫是马尾松林的主要害虫,危害很大,实施有效监测是对其进行防治的关键。以福建沙县为研究区,以SPOT-5遥感影像为数据源,基于短波红外波段、红光波段和近红外波段等构造归一化植被指数等光谱特性指数,基于灰度共生矩阵法提取均值等影像纹理特征,引入坡度、坡向等地形及林木因子构建马尾松毛虫虫害遥感监测指标,通过岭回归分析建立虫情级数估测模型进而反演,进行虫害监测信息提取,并与单纯基于光谱特性指数指标的虫害信息提取方法进行比较。结果表明:1)综合考虑影像的光谱特性、纹理特征与地形及林木因子影响进行虫害监测,显著增强了虫害的光谱响应能力,相比于单纯基于光谱特性的虫害监测信息提取,其总精度提高了14.28%;2)这些为林业生产实践提供技术支撑与理论借鉴,同时也在一定程度上推动遥感监测森林病虫害的理论认识与信息提取方法。
Dendrolimus punctatus is a major pest of Pinus massoniana forests and is very harmful, effective monitoring is the key step for controlling Dendrolimus punctatus. Shaxian county of Fujian Province was chosen as the study area and SPOT-5 Remote Sensing Image as the data sources. Normalized vegetation indexes(other spectral characteristics indexes) were generated based on shortwave infrared band, red band and near infrared band. The image texture features such as mean values were extracted based on gray level cooccurrence matrix method. The remote sensing monitoring index system of D.punctatus insect pests was constructed by introducing topographic factors such as slope, slope direction and forest tree factors. And the insect pest level estimation model was established by ridge regression analysis, and then the inversion was carried out. With this model, the insect pest monitoring information was extracted and compared with that of the pest information extraction method based solely on spectral characteristic index. The results show that 1)the method of comprehensively considering spectrum characteristic, texture features, terrain and forest factors had better effects than the method only based on spectrum characteristic, which highly enhanced the D. punctatus damage spectral response and the total accuracy was Increased by 14.28%; 2) These methods provide a technical support and theoretical reference for forestry production practices, and at the same time promote the theoretical understanding and information extraction methods of remote sensing monitoring of forest pests and diseases.
Dendrolimus punctatus is a major pest of Pinus massoniana forests and is very harmful, effective monitoring is the key step for controlling Dendrolimus punctatus. Shaxian county of Fujian Province was chosen as the study area and SPOT-5 Remote Sensing Image as the data sources. Normalized vegetation indexes(other spectral characteristics indexes) were generated based on shortwave infrared band, red band and near infrared band. The image texture features such as mean values were extracted based on gray level cooccurrence matrix method. The remote sensing monitoring index system of D.punctatus insect pests was constructed by introducing topographic factors such as slope, slope direction and forest tree factors. And the insect pest level estimation model was established by ridge regression analysis, and then the inversion was carried out. With this model, the insect pest monitoring information was extracted and compared with that of the pest information extraction method based solely on spectral characteristic index. The results show that 1)the method of comprehensively considering spectrum characteristic, texture features, terrain and forest factors had better effects than the method only based on spectrum characteristic, which highly enhanced the D. punctatus damage spectral response and the total accuracy was Increased by 14.28%; 2) These methods provide a technical support and theoretical reference for forestry production practices, and at the same time promote the theoretical understanding and information extraction methods of remote sensing monitoring of forest pests and diseases.
引文
[1]庞正轰.马尾松毛虫灾害预测预报综合技术研究[D].北京:北京林业大学,2004.
[2]S.P.EKSTRAND.Detection of Moderrate Damage on Noway Spruce Using Landsat TM and Digital Stand Data[J].IEEETransaction on Geoscience and Remote Sensing,1990,28(4):685-692.
[3]R.H.FRASTER.Mapping insect induced tree defoliation and mortality using coarse spatial resolution satellite imagery[J].International Journal of Remoteing Sensing,2005,26(1):193-200.
[4]M.D.GILLS,R.D.PICK,D.G.1eckie.Satellite imagery assists in the assessment of hail damage for salvage harvest[J].The Forestry Chronicle,1990,32(10):463-468.
[5]R.F.NELSON.Detecting forest canopy change due to insect activity using Landsat MSS[J].PE&RS,1983(49):1303-1314.
[6]A.N.RENCZ,J.NEMETH.Defection of Mountain Pine Beetle Infestation Using Landsat Mss and Simulated Thematic Mapper Data[J].Canadian Journal of Remote Sensing,1985(11):50-58.
[7]P.E.JORIA,S.C.AHEARN.A Comparision of the SPOT and Landsat Thematic Mapper Satellite Systems for Detecting Gypsy Moth Defoliation in Michigan[J].Photogrammetric Engineering&Remote Sensing,1991,57(12):1605-1612.
[8]V.I.KHARUK.NOAA/AVHRR satellite detection of Siberian silkmoth outbreaks in eastern Siberia[J].INT J Remote Sensing,2004,25(24):5543-5555.
[9]V.C.RADELOFF,D.J.MLANDENOF F,M.S.BOYCE.Detecting jack pine budworm defoliation using spectral mixture analysis:sparating effects from determinants[J].Remote Sense Envion,1999(69):156-169.
[10]K.P.PRICE,M.E.JAKUBAUSKAS.Spectral retrogression and insect damage in lodgepole pine successional forests[J].Int JRemote Dens,1998,19(8):1627-1632.
[11]D.D.ROYLE,R.G.LATHROP.Monitoring hemlock forest health in new jersey using landsat TM data and change detection techniques[J].For Sci,1997,43(3):327-335.
[12]S.E.FRANKLIN.Classification of hemlock looper defoliation using SPOT HRV imagery[J].Can J Remote Sensing,1989(20):37-48.
[13]Y.MUKAI,T.SUGIMURA,H.WATANABE,et al.Extraction of areas infested by pine bark beetle using Landsat MSS data[J].C,1987(54):77-81.
[14]武红敢,乔彦友,黄建文.利用陆地卫星TM数据评估森林病虫害[J].遥感技术与应用,1994(4):46-51.
[15]武红敢,石进.松毛虫灾害的TM影像监测技术[J].遥感学报,2004(2):172-177.
[16]刘清旺,武红敢,石进,等.基于TM影像的森林病虫灾害遥感监测系统[J].遥感应用,2007(2):46-49.
[17]刘志明,晏明,张旭东,等.用气象卫星监测大范围森林虫害方法研究[J].自然灾害学报,2002,11(3):109-114.
[18]亓兴兰,刘健,陈国荣,等.应用MODIS遥感数据监测马尾松毛虫虫害研究[J].西南林学院学报,2010,30(1):42-46.
[19]戴昌达,雷莉萍,胡德永,等.卫星遥感监测松毛虫灾害[J].遥感信息,1991(3):32-34.
[20]亓兴兰,胡宗庆,刘健,等.基于光谱特征的SPOT-5影像马尾松毛虫虫害信息提取[J].东北林业大学学报,2012,40(5):131-136.
[21]郭志华,肖文发,张真,等.RS在森林病虫害监测研究中的应用[J].自然灾害学报,2003,12(4):73-81.
[22]王蕾,黄华国,张晓丽,等.3S技术在森林虫害动态监测中的应用研究[J].世界林业研究,2005,18(2):51-56.
[23]D.HANKERSON,G.A.HARRIS,P.D.JOHNSON.Introduction to information theory and data compression[M].New York:CRCPress,1997.
[24]谭永生.像素级中高分辨率遥感影像融合研究[D].杭州:浙江大学,2007.
[25]亓兴兰.SPOT-5遥感影像马尾松毛虫虫害信息提取技术研究[D].福州:福建农林大学,2011.
[26]国家林业局.森林病虫害防治知识问答[M].北京:中国林业出版社,1999:2-5.
[27]汪康宁,吕杰,李崇贵.基于多尺度遥感影像纹理特征的森林蓄积量反演[J].中南林业科技大学学报,2017,37(11):84-89.
[28]黄二辉,潘德炉,李淑菁,等.水下剖面光谱原始数据异常值的判断方法[J].海洋学研究,2006,24(1):91-96.
[29]任若恩,王惠文.多元统计数据分析-理论、方法、实例[M].北京:国防工业出版社,1997:92-111.
[30]粟丽,赵伟,王志福.在解决多重共线性问题上岭回归法比LS法的优越性[J].渤海大学学报(自然科学版),2006,27(2):62-68.
[2]S.P.EKSTRAND.Detection of Moderrate Damage on Noway Spruce Using Landsat TM and Digital Stand Data[J].IEEETransaction on Geoscience and Remote Sensing,1990,28(4):685-692.
[3]R.H.FRASTER.Mapping insect induced tree defoliation and mortality using coarse spatial resolution satellite imagery[J].International Journal of Remoteing Sensing,2005,26(1):193-200.
[4]M.D.GILLS,R.D.PICK,D.G.1eckie.Satellite imagery assists in the assessment of hail damage for salvage harvest[J].The Forestry Chronicle,1990,32(10):463-468.
[5]R.F.NELSON.Detecting forest canopy change due to insect activity using Landsat MSS[J].PE&RS,1983(49):1303-1314.
[6]A.N.RENCZ,J.NEMETH.Defection of Mountain Pine Beetle Infestation Using Landsat Mss and Simulated Thematic Mapper Data[J].Canadian Journal of Remote Sensing,1985(11):50-58.
[7]P.E.JORIA,S.C.AHEARN.A Comparision of the SPOT and Landsat Thematic Mapper Satellite Systems for Detecting Gypsy Moth Defoliation in Michigan[J].Photogrammetric Engineering&Remote Sensing,1991,57(12):1605-1612.
[8]V.I.KHARUK.NOAA/AVHRR satellite detection of Siberian silkmoth outbreaks in eastern Siberia[J].INT J Remote Sensing,2004,25(24):5543-5555.
[9]V.C.RADELOFF,D.J.MLANDENOF F,M.S.BOYCE.Detecting jack pine budworm defoliation using spectral mixture analysis:sparating effects from determinants[J].Remote Sense Envion,1999(69):156-169.
[10]K.P.PRICE,M.E.JAKUBAUSKAS.Spectral retrogression and insect damage in lodgepole pine successional forests[J].Int JRemote Dens,1998,19(8):1627-1632.
[11]D.D.ROYLE,R.G.LATHROP.Monitoring hemlock forest health in new jersey using landsat TM data and change detection techniques[J].For Sci,1997,43(3):327-335.
[12]S.E.FRANKLIN.Classification of hemlock looper defoliation using SPOT HRV imagery[J].Can J Remote Sensing,1989(20):37-48.
[13]Y.MUKAI,T.SUGIMURA,H.WATANABE,et al.Extraction of areas infested by pine bark beetle using Landsat MSS data[J].C,1987(54):77-81.
[14]武红敢,乔彦友,黄建文.利用陆地卫星TM数据评估森林病虫害[J].遥感技术与应用,1994(4):46-51.
[15]武红敢,石进.松毛虫灾害的TM影像监测技术[J].遥感学报,2004(2):172-177.
[16]刘清旺,武红敢,石进,等.基于TM影像的森林病虫灾害遥感监测系统[J].遥感应用,2007(2):46-49.
[17]刘志明,晏明,张旭东,等.用气象卫星监测大范围森林虫害方法研究[J].自然灾害学报,2002,11(3):109-114.
[18]亓兴兰,刘健,陈国荣,等.应用MODIS遥感数据监测马尾松毛虫虫害研究[J].西南林学院学报,2010,30(1):42-46.
[19]戴昌达,雷莉萍,胡德永,等.卫星遥感监测松毛虫灾害[J].遥感信息,1991(3):32-34.
[20]亓兴兰,胡宗庆,刘健,等.基于光谱特征的SPOT-5影像马尾松毛虫虫害信息提取[J].东北林业大学学报,2012,40(5):131-136.
[21]郭志华,肖文发,张真,等.RS在森林病虫害监测研究中的应用[J].自然灾害学报,2003,12(4):73-81.
[22]王蕾,黄华国,张晓丽,等.3S技术在森林虫害动态监测中的应用研究[J].世界林业研究,2005,18(2):51-56.
[23]D.HANKERSON,G.A.HARRIS,P.D.JOHNSON.Introduction to information theory and data compression[M].New York:CRCPress,1997.
[24]谭永生.像素级中高分辨率遥感影像融合研究[D].杭州:浙江大学,2007.
[25]亓兴兰.SPOT-5遥感影像马尾松毛虫虫害信息提取技术研究[D].福州:福建农林大学,2011.
[26]国家林业局.森林病虫害防治知识问答[M].北京:中国林业出版社,1999:2-5.
[27]汪康宁,吕杰,李崇贵.基于多尺度遥感影像纹理特征的森林蓄积量反演[J].中南林业科技大学学报,2017,37(11):84-89.
[28]黄二辉,潘德炉,李淑菁,等.水下剖面光谱原始数据异常值的判断方法[J].海洋学研究,2006,24(1):91-96.
[29]任若恩,王惠文.多元统计数据分析-理论、方法、实例[M].北京:国防工业出版社,1997:92-111.
[30]粟丽,赵伟,王志福.在解决多重共线性问题上岭回归法比LS法的优越性[J].渤海大学学报(自然科学版),2006,27(2):62-68.