针对目前合成孔径雷达(SAR)与可见光图像融合结果目标信息缺失、对比度不高的缺点，提出了一种基于纹理分割和top-hat变换的图像增强融合算法。将SAR图像灰度共生矩阵的熵纹理特征图进行阈值分割，提取SAR图像的感兴趣区域（ROI）；并对SAR和可见光图像进行非下采样Contourlet变换(NSCT)分解，低频系数采用基于区域的融合规则，在感兴趣区域内选择SAR的低频系数。对低频系数进行top-hat变换得到显著化的图像亮、暗细节特征，并加入到低频系数上形成低频合成系数；高频子带系数采用局部方向信息熵显著性因子取大的融合规则；对融合系数进行NSCT逆变换得到最终的融合图像。实验证明了本算法的有效性。

To overcome the disadvantages of low contrast and missing target information for the existing synthetic aperture radar (SAR) and visible image fusion methods, an image enhanced fusion algorithm based on texture segmentation and top-hat transformation is proposed. Entropy texture image which is generated by gray level co-occurrence matrix of SAR image is segmented by threshold, and the region of (ROI) interest of SAR is extracted. The SAR and optical images are decomposed by the non-subsampled contourlet transform (NSCT). A region fusion rule is introduced to the low-frequency coefficient, and low-frequency coefficient of SAR is chosen in the region of interest. The significant bright and dark image detail features are extracted by top-hat transformation and the low-frequency synthetic coefficient is obtained through adding the above bright and dark features into low-frequency coefficient. High-frequency subband coefficients are fused by selecting maximum significant factor of local directional entropy. The fused image is obtained by the NSCT inverse transformation of the fused coefficient. The experiment results testify the validity of the proposed image fusion algorithm.