针对多云雾地区高时空分辨率数据缺乏现状,提出了一套区域尺度高时空分辨率植被覆盖度数据构建方法。首先,通过时空适应反射率融合模型(STARFM)有效地将TM的较高空间分辨率与MODIS的高时间分辨率融合在一起,构建了研究区植被生长峰值阶段的NDVI数据;然后,以植被生长峰值阶段的NDVI为输入,基于地表覆被类型,综合应用等密度和非密度亚像元模型对研究区的植被覆盖度进行估算。结果表明:①即使数据源存在大量的云雾,且存在一定的时相差异,研究区植被覆盖度的估算结果过渡自然,不存在明显的不接边效应;②以植被生长峰值阶段的NDVI数据为输入进行植被覆盖度估算,有效拉开了同一地表覆被类型不同覆盖度像元的NDVI梯度,提高了亚像元估算模型对输入数据的抗扰动性;③基于地表覆被类型,应用亚像元混合模型,能够提高植被覆盖度的估算精度。经野外实测数据验证,总体约85%的估算精度表明,针对高时空分辨率遥感数据缺乏的多云雾区域,本研究提出的方法能够实现区域尺度植被覆盖度数据的构建。

Focusing on the cloudy and foggy region lacked of remote sensing data with highspatial and temporal resolution,a method of constructing vegetationfractional coveragewith high spatial and temporal resolution,on region scale has been proposed,in this paper.First,normalized difference vegetation index(NDVI) data with higher spatial andtemporal resolution was constructed by combining advantages both of TM and MODIS usingspatial and temporal adaptive reflectance fusion model(STARFM).Then,based on the land coverage typedata and NDVI data in the peak stage of vegetation growth,the vegetation fractional coverageof study area was estimated by density sub-pixel model andnon-density sub-pixel model.The result shows that:①Even the data sources with a lot of snow and cloud or shoot at different times,in the estimated vegetation fractional coverage image,the color of the area covered by cloud or it’s shade is consistent with the color of area uncontaminated;②The normalizeddifference vegetation index(NDVI)data,in the peak stage of vegetation growth,as input dataimproved the anti-disturbance to input data of sub-pixel mixed model estimates of vegetationcoverage by maximum the NDVI difference of same vegetation with differentvegetation fractional coverage;③Based on land coverage using sub-pixel model can improve the accuracy ofestimating vegetation fractional coverage.Validated by the data measured in the field,the accuracy of estimated vegetation fractional coverage isabout 85%,which suggest that it is viable to estimate vegetation fractional coverage in large regions,especially lacking of remote sensing data with high spatial and temporal resolution.