【摘要】 提出一个应用方便的标准化差值植被指数（ＮＤＶＩ）的大气纠正方法。在大气阻抗植被指数（ＡＲＶＩ）的基础上，运用大气下向光谱的同步实例值以及大气辐射传输方程，得到纠正ＮＤＶＩ的关键参数γ↑。改进了ＡＲＶＩ的γ↑始终等于１ 的计算方法。同时也不必采用５ｓ模型进行预处理。所得新的抗大气影响植被指数（ＩＡＶＩ）为：ρｎｉｒ－ 〔ρｒ－ γ↑（ρｂ－ ρｒ）〕ρｎｉｒ＋ 〔ρｒ＋ γ↑（ρｂ－ ρｒ）〕。式中的γ↑值的变化范围可从０．６５ 到１．２１ 之间。对于混浊大气γ↑值随观测角变化更大。还提供了一组根据气象资料计算γ↑值的近似公式。根据应用表明ＩＡＶＩ比ＮＤＶＩ有明显的抗大气影响效果更多还原

【Abstract】 Estimation of vegetation information at the regional scales using remote sensing data has been a subject of considerable research and development in view of the importance of land atmosphere interactions in global climate change.The normalized difference vegetation index (NDVI) approach has been a popular method.Many experiments on the ground had indicated that NDVI was closely correlated to the leaf area index (LAI) and photosynthetically active radiation (PAR). In order to determine LAI and PAR accurately in a region by means of NDVI from available satellite inages,atmospheric effects must be largely elucidated. Although substantial progress has been made to improve the NDVI method,removal of the atmospheric effects has not been completely satisfactory.The atmospheric effect on NDVI has been contemplated in the experiments conducted at Yucheng Experimental Comprehensive Station,China.Kaufman and his colleagues also had verified the atmospheric effect on NDVI even the effect still existed after using 5s algorithm to remove rayleigh scattering and ozone absorption.For this reason they presented an atmosphere resistant vegetation index (ARVI) which was based on correlation between blue and red wave bands of the atmosphere.However according to our practice the authors found that the improvement using ARVI with a constant “1” correction factor was not sufficient for a number of atmospheric conditions, and it was not convenient to use 5s algorithm that many atmosphere parameters were still required. The authors has developed a new atmospheric effect resistant vegetation index (IAVI),which removed the atmospheric effect for determination of the vegetation index using the atmosphere downward path radiation measured with a spectrum radiometer and using information of remote sensing images.The approach avoided using the inconvenient 5s algorism. The process of obtaining the IAVI summarized as follows: 1.Determine the downward atmosphere path radiance on the ground using the spectrum radiometer with a simultaneous measurement as the satellite passed. 2.Establish the mathematic relationship between the downward and upward path radiation based on a radiation transfer model. 3.Determine the upward path radiance using the downward path radiance and the function or look up charts established in step 2. 4.Calculate the coefficient γ ↑ using the upward path radiance in the red and blue wave bands. 5.Obtain the new IAVI for various viewing angles. The authors also proposed the values for calculating IAVI in absence of the gound based spectrum radiometer measurements to approximately correct the atmosphere effect.For summer condition: γ ↑=0.656-0.006 sinθ+0.124 sin 2θ for rural area,visibility>22 km γ ↑=0.656-0.018 sinθ+0.106 sin 2θ for urban area,visibility>22 km γ ↑=0.722-0.011 sinθ+0.165 sin 2θ for rural area, 7 km<visibility<20 km γ ↑=0.646-0.002 sinθ+0.140 sin 2θ for urban area, 7 km<visibility<20 km γ ↑=0.788+0.015 sinθ+0.206 sin 2θ for rural area, visibility<5 km γ ↑=0.643+0.007 sinθ+0.174 sin 2θ for urban area, visibility<5 km For winter condition: γ ↑=0.677+0.019 sinθ+0.200 sin 2θ for rural area,visibility>22 km γ ↑=0.642-0.009 sinθ+0.196 sin 2θ for urban area,visibility>22 km γ ↑=0.785+0.040 sinθ+0.237 sin 2θ for rural area,7 km<visibility<20 km γ ↑=0.664+0.021 sinθ+0.223 sin 2θ for urban area,7 km<visibility<20 km γ ↑=0.893+0.060 sinθ+0.273 sin 2θ for rural area,visibility<5 km γ ↑=0.686+0.051 sinθ+0.249 sin 2θ for urban area,visibility<5 km The correction factor γ ↑was a crucial parameter for upward path radiance.It varied with atmosphere aerosol condition and viewing angles and was deviated from 1.Based on the above analysis,the range of γ ↑ was from 0.65 to 1.21.For hazy atmosphere conditions it varied more significantly with aerosol type and viewing angles than clear condition.The percentage of error of vegetation index for the two cases studied were less than 4% as compared with 16% and 31% using NDVI,and 19% u更多还原