结合雷达影像纹理特征的作物叶面积指数估测

The Estimation of Crop Leaf Area Index in Consideration of Texture Characteristics of SAR

研究了结合雷达影像纹理特征进行作物叶面积指数(LAI)估测的可行性,分析了作物LAI与多极化雷达纹理特征的相关性。将ENVISATASAR影像6种纹理特征与实测玉米的LAI进行相关分析发现,HH极化影像的灰度共生矩阵纹理特征与玉米LAI的相关性较VV极化的高;HH极化影像的对比度、异质性,VV极化影像的偏度、均质性等纹理特征与实测LAI均有较大的相关性。分别对两种极化影像雷达波散射强度及纹理特征与实测LAI进行多元回归分析,发现HH极化方式的相关系数达到0.68、VV极化的为0.87。说明结合纹理特征的雷达估测作物LAI方法具有一定的可行性。

The authors studied the feasibility of estimating Leaf Area Index （LAI） of the crop by using intensity and texture characteristics of SAR, and analyzed the texture characteristics of SAR which have relatively high correlation with LAI. In this study, six texture characteristics calculated from ENVISAT - ASAR image were selected and compared with measured LAI of the corn. The results show that the texture characteristics of HH polarization for gray level co - occurrence matrix have higher correlation with the LAI of corn than those of W polarization. Dissimilarity of HH polarization and skewness and homogeneity of VV polarization are significantly related to LAI. In combination with backscattering coefficient, multiple regressions of two formulae were computed respectively, and the correlation coefficients are 0.68 for HH polarization and 0.87 for VV polarization. It is thus held that the methods discussed in this paper have potential application values in the estimation of the crop Leaf Area Index.