利用R-Shiny架构的植被物候参数分析系统设计与实现
|
摘要
植被物候作为反映植被与气候变化关系的重要参量,具有重要的研究意义。本文基于R语言分布式架构Shiny构建了植被物候参数分析系统,可实现站点分布可视化、感兴趣区(Region of Interest,ROI)选取与绘制、植被指数计算与可视化、数据过滤、生长曲线轨迹拟合与物候参数提取等功能模块。用户可提取不同植被类型数码相机时间序列的植被指数,并用max方法进行平滑与去噪处理,然后选择合适的方法组合拟合植被群落季相变化轨迹,最终提取较为精确的关键物候参数。林地数据系统测试结果表明:1)相对绿度指数GI比其他相对植被指数和单波段的亮度值振幅明显,其时间序列可表征植被实际生长轨迹; 2)不同拟合方法与提取方法的组合效果不同,如klosterman与klosterman方法组合适合林地类型的植被物候参数提取,用户可综合均方根误差与季节群相变化轨迹结果,筛选出适合所选植被物候数据的拟合与物候参数提取方法组合。
Vegetation phenology has important research significance as an important parameter reflecting the relationship between vegetation and climate change. In this paper,a web system of vegetation phenology parameters extraction and analysis is constructed based on Shiny,which can realize functional modules such as site distribution visualization,ROI(Region of Interest) selection,vegetation index calculation and visualization,data filtering,growth curve trajectory fitting and phenological parameters extraction. The user can extract the vegetation index time series of digital camera with different vegetation types and use " max" method for smoothing and denoising. Then we can select the appropriate method to fit the seasonal variation trajectory of vegetation communities and finally extract the more accurate key phenological parameters. The system test results with woodland data show that:(1) The amplitude of relative green index(GI) is obviously larger than those of other relative vegetation indices and single band value,so its time series can characterize the actual growth trajectory of vegetation;(2) Different combinations of fitting methods and extraction methods have different effects. For example,the combination of klosterman and klosterman is suitable for the extraction of the vegetation phenology parameters of the woodland type. The user can synthesize the RMSE and the seasonal variation trajectory results to select the fitting method and phenological parameter extraction method suitable for the selected vegetation phenology data.
Vegetation phenology has important research significance as an important parameter reflecting the relationship between vegetation and climate change. In this paper,a web system of vegetation phenology parameters extraction and analysis is constructed based on Shiny,which can realize functional modules such as site distribution visualization,ROI(Region of Interest) selection,vegetation index calculation and visualization,data filtering,growth curve trajectory fitting and phenological parameters extraction. The user can extract the vegetation index time series of digital camera with different vegetation types and use " max" method for smoothing and denoising. Then we can select the appropriate method to fit the seasonal variation trajectory of vegetation communities and finally extract the more accurate key phenological parameters. The system test results with woodland data show that:(1) The amplitude of relative green index(GI) is obviously larger than those of other relative vegetation indices and single band value,so its time series can characterize the actual growth trajectory of vegetation;(2) Different combinations of fitting methods and extraction methods have different effects. For example,the combination of klosterman and klosterman is suitable for the extraction of the vegetation phenology parameters of the woodland type. The user can synthesize the RMSE and the seasonal variation trajectory results to select the fitting method and phenological parameter extraction method suitable for the selected vegetation phenology data.
引文
[1] White M A,Thomton P E,Running S W. A continentalphenology model for monitoring vegetation responses to in-terannual climatic variability[J]. Global BiogeochemicalCycles,1997,11(2):217-234.
[2] Zhang X Y,Friedl M A,Schaaf C B,et al.Monitoring vege-tation phenology using MODIS[J].Remote Sensing of En-vironment,2003,84(3):471-475.
[3]刘玲玲,刘良云,胡勇.基于AVHRR和MODIS数据的全球植被物候比较分析[J].遥感技术与应用,2012,27(5):754-762.
[4]康峻,侯学会,牛铮,等.基于拟合物候参数的植被遥感决策树分类[J].农业工程学报,2014,30(9):148-156.
[5]廖韵,董奎,付静静,等.都江堰市银杏物候时序特征的数码相机监测与分析[J].四川林业科技,2016(3):54-61.
[6]周惠慧,付东杰,张立福,等.基于数字相机的草地物候模拟及其与气象因子关系的研究[J].遥感技术与应用,2016,31(5):966-974.
[7] Filippa G,Cremonese E,Galvagno M,et al.Five years ofphenological monitoring in a mountain grassland:inter-annual patterns and evaluation of the sampling protocol[J]. International Journal of Biometeorology,2015,59(12):1 927-1 937.
[8] Beck P S A,Atzberger C,Hgda K A,et al.Improved mo-nitoring of vegetation dynamics at very high latitudes:anew method using MODIS NDVI[J]. Remote Sensing ofEnvironment,2006,100(3):321-334.
[9] Gu L,Post W M,Baldocchi D D,et al.Characterizing theseasonal dynamics of plant community photosynthesisacross a range of vegetation types[M].Phenology of Eco-system Processes.Springer New York,2009.
[10] Elmore A J,Guinn S M,Minsley B J,et al.Landscape con-trols on the timing of spring,autumn,and growing seasonlength in mid-Atlantic forests[J].Global Change Biology,2012,18(2):656-674.
[11] Klosterman S T,Hufkens K,Gray J M,et al.Evaluating re-mote sensing of deciduous forest phenology at multiple spa-tial scales using PhenoCam imagery[J]. Biogeosciences,2014,11(16):4 305-4 320.
[2] Zhang X Y,Friedl M A,Schaaf C B,et al.Monitoring vege-tation phenology using MODIS[J].Remote Sensing of En-vironment,2003,84(3):471-475.
[3]刘玲玲,刘良云,胡勇.基于AVHRR和MODIS数据的全球植被物候比较分析[J].遥感技术与应用,2012,27(5):754-762.
[4]康峻,侯学会,牛铮,等.基于拟合物候参数的植被遥感决策树分类[J].农业工程学报,2014,30(9):148-156.
[5]廖韵,董奎,付静静,等.都江堰市银杏物候时序特征的数码相机监测与分析[J].四川林业科技,2016(3):54-61.
[6]周惠慧,付东杰,张立福,等.基于数字相机的草地物候模拟及其与气象因子关系的研究[J].遥感技术与应用,2016,31(5):966-974.
[7] Filippa G,Cremonese E,Galvagno M,et al.Five years ofphenological monitoring in a mountain grassland:inter-annual patterns and evaluation of the sampling protocol[J]. International Journal of Biometeorology,2015,59(12):1 927-1 937.
[8] Beck P S A,Atzberger C,Hgda K A,et al.Improved mo-nitoring of vegetation dynamics at very high latitudes:anew method using MODIS NDVI[J]. Remote Sensing ofEnvironment,2006,100(3):321-334.
[9] Gu L,Post W M,Baldocchi D D,et al.Characterizing theseasonal dynamics of plant community photosynthesisacross a range of vegetation types[M].Phenology of Eco-system Processes.Springer New York,2009.
[10] Elmore A J,Guinn S M,Minsley B J,et al.Landscape con-trols on the timing of spring,autumn,and growing seasonlength in mid-Atlantic forests[J].Global Change Biology,2012,18(2):656-674.
[11] Klosterman S T,Hufkens K,Gray J M,et al.Evaluating re-mote sensing of deciduous forest phenology at multiple spa-tial scales using PhenoCam imagery[J]. Biogeosciences,2014,11(16):4 305-4 320.