论文部分内容阅读
随着多空间分辨率、多光谱分辨率、多传感器遥感数据日益增多,数据融合技术已经在地学领域得到广泛应用。以SAR和SPOT多光谱影像为数据源,通过几何校正、辐射定标、噪声去除等预处理,采用PCAI、HS、BT、HPF以及WT融合方法进行融合试验,选取标准差、信息熵以及相关系数对融合结果进行图像质量评价。结果表明:基于结构信息变换的HPF和WT方法,纹理结构信息保持较好,线性地物特征尤为突出;基于统计信息变换的PCA方法,不仅能较好保持多光谱信息,而且也保持了SAR影像纹理结构信息,信息熵指数最大;基于彩色变换的IHS和BT方法,虽然视觉效果较理想,但是色彩及光谱失真现象严重,信息熵值最小。
With the multi-spatial resolution, multi-spectral resolution and multi-sensor remote sensing data increasing, data fusion technology has been widely used in geosciences. The fusion experiments of PCAI, HS, BT, HPF and WT fusion were carried out by using SAR and SPOT multispectral images as data sources. The standard deviation, the entropy of information and the correlation coefficients were selected by pretreatment such as geometric correction, radiation calibration and noise removal. The fusion results were evaluated for image quality. The results show that, based on the HPF and WT methods of structural information transformation, the texture structure information is well preserved and the features of linear features are particularly prominent. The PCA method based on statistical information transformation can not only better maintain the multi-spectral information, but also keep the SAR image Texture structure information and information entropy index are the largest. The IHS and BT methods based on color transform, although the visual effect is better, but the color and spectral distortion are serious, and the information entropy is the smallest.