论文部分内容阅读
本文提出了一种新的混合像元聚集指数计算方法。该方法基于高分辨率遥感数据,采用方向孔隙率公式和线性混合模型相结合的方法,详细描述了混合像元内部由于植被覆盖度(FVC)、端元聚集指数以及叶倾角分布等引起的尺度差异。利用模拟数据进行初步敏感性分析,将计算得到的混合像元聚集指数与所占面积比例最大的端元聚集指数进行比较,结果表明端元聚集指数的不均一性对计算结果影响最大,由此得到的混合像元聚集指数较最大面积端元聚集指数下降了55%;植被覆盖度的不均一性影响次之,可使聚集指数降低43%;植被叶倾角分布(G函数)的空间异质性影响最小,约12%。敏感性分析的结果同时也证明了对于空间异质性较强的混合像元进行尺度差异修正的必要性。利用本文中提出的混合像元聚集指数方法有望在提高低分辨率叶面积指数反演中发挥重要作用。
In this paper, a new method for calculating the clustering index of mixed pixels is proposed. Based on the high-resolution remote sensing data and the combination of the directional porosity formula and the linear mixed model, the method described in detail the scale (FVC), the clustering index of end elements and the distribution of the leaf inclination angle in the mixed pixel, difference. The initial sensitivities of simulated data were compared and the aggregated index of aggregated pixels was compared with that of the largest aggregated area. The results showed that the heterogeneity of aggregated index of end elements had the most significant effect on the calculated results. Thus, The aggregated index of mixed pixels decreased by 55% compared with that of the largest area, followed by the heterogeneity of vegetation coverage, which reduced the aggregation index by 43%. Spatial heterogeneity of vegetation leaf inclination distribution (G function) The least impact, about 12%. The result of sensitivity analysis also proves the necessity of scaling differences for mixed pixels with strong spatial heterogeneity. The hybrid pixel clustering index proposed in this paper is expected to play an important role in improving low-resolution leaf area index inversion.