An Adaptive Multiscale Gaussian Co-Occurrence Filtering Decomposition Method for Multispectral and SAR Image Fusion

Abstract

Spectral information and backscatter information are both exclusively important bases for land cover classification, and these two kinds of information are found in multispectral images and SAR images, respectively. Therefore, the fusion of complementary information of multispectral and SAR images can effectively improve land cover classification accuracy. However, the existing fusion methods of multispectral and SAR images generally have some problems, such as insensitivity to edge information, serious interference by speckle noise, and unreasonable settings of fusion rules, which lead to unsatisfactory results of land cover classification. To solve this issue, a fusion method based on adaptive multiscale Gaussian co-occurrence filtering decomposition is proposed. Firstly, parameters controlling Gaussian distribution scale in co-occurrence filter are adaptive according to pixel intensity skewness. Gaussian filtering and adaptive co-occurrence filtering are applied to the original image to smooth out speckle noise and interference edges within the textures while preserving edge information between textures. Secondly, difference calculation among the original image, Gaussian filtered image and adaptive co-occurrence filtered image is carried out to realize the decomposition of detail information, edge information and basic information. Compared with the decomposition method that only decomposes detail information and basic information, it can effectively preserve boundary structure between the features. The image decomposition is extended to multi-scale space for multi-layer fusion of image features at the same time. Finally, according to the decomposition scale, the decomposition results are divided into three layers, and corresponding rules of fusion are formulated based on their feature information, so as to generate fusion image with low noise interference, clear boundary and homogeneous pixel convergence. Experimental results show that the proposed method generally performs the best in eight evaluation indexes compared with ten other methods. The overall accuracy, average accuracy and Kappa coefficient of land cover classification are increased by 7.674%, 6.776% and 0.098 respectively compared with those of the original multispectral image in Area 1, and by 6.904%, 7.649% and 0.089 respectively compared with those of the original multispectral image in Area 2.

Topics & Concepts