【摘要】 针对人眼对于外界信息的摄取会进行过滤,对于一幅既定的场景,会将其分为目标和背景两部分,对于目标信息的获取会希望更加详细,背景信息没有过多要求的特点,本文研究了联合图像分类的图像融合算法。对给定的待融合图像进行NSCT(Non-subsampled contourlet transform)变换,在NSCT变换域内提取特征,利用K-Means方法将图像分为目标和背景两部分,然后对背景和目标的低频信息采用均值准则、目标的高频信息采用区域能量加权平均的准则进行融合;同时为了降低计算复杂度,将压缩感知应用于图像分类过程中,实验结果验证了本文算法的优越性。更多还原

【Abstract】 For a given scene, human’s eye will filter external information and divide the image into two parts: the target and the background. The target information acquisition will be more detailed, while the background information may be ignored to some extent. A novel image fusion algorithm combining with image classification is proposed in this paper according to the image characteristics. The images will be transformed in Non-subsampled contourlet transform(NSCT) domain and will be classified into object and background parts by K-Means. Then the low-frequency information of background and target is fused by the mean value criterion, and the high-frequency information of target is fused by the weighted average criterion of regional energy. Meanwhile, to reduce the computational complexity, the Sensing Compressive(CS) is applied to image classification processing. The advantages of the proposed method are proved by the experimental results.更多还原