土地覆盖分类影像融合与精度评价研究

Research on Image Fusion and Precision Evaluation of Land Cover Classification

影像融合技术可以使遥感影像具有高光谱和高空间分辨率的效果,实现不同空间、光谱、时间等多种分辨率的信息资源互补,从而提高图像的空间分辨率,提高图像的几何精度.文章利用ERDAS软件,对遥感影像数据进行融合,采用乘积变换、PCA变换、Brocey变换、小波变换等遥感影像融合方法对多光谱与全色影像进行融合和土地覆盖分类研究.通过结合图像的光谱统计参数和融合图像的分类精度,对这些方法的分类精度进行评价.这4种方法对于原始影像分类精度,均有不同程度的提高.而小波变换所得融合影像与原多光谱影像的相关系数最大,均方差、平均梯度和信息熵最大,偏差指数最小,影像所含信息量最多;在光谱特性、图像清晰度、对于空间细节信息的表现能力上其它三种方法都好,所得融合影像的分类精度也是最高的.小波变换更适合融合影像的土地覆盖分类研究.

Image fusion technology can make the remote sensing images with high spectral and high spatial resolution effect, different space, time , and other complementary spectral resolution of the information resources, so as to improve the spatial resolution o f the image and improve the geometry precision of image. Using ERDAS software and the remote sensing image data fusion , the product （multiplicative ） trans fo rm, PCA transform , Brocey transform, wavelet transform （ wavelet） etc. Remote sensing image fusion method of multi spectral and panchromatic image fusion and land cover classification. The classification accuracy of these methods is evaluated by combining the spectral stat ist ical parameters of the image and the classification accuracy of the fused image. The results show that these 4 methods have d iffe re nt degrees o f improvement in classification accuracy of original image. And wavelet transform （ wavelet） income fusion image and or igin al multi spectral image correlation coef ficient is the biggest, are variance, average gradient and the information entropy,the deviation index minimum image contains informat ion most; in the spectral charac teristics, image clarity,detail spatial informat ion table now ability on the other three methods are good fo r, fused images obtained by the classif icat ion accuracy is the highest.The wavelet transform （Wavelet） is more suitable for the land cover classification of image fusion.