摘要
目的第二代curvelet变换与图像区域能量相结合的方法实现多聚焦图像融合。方法对两幅多聚焦图像,分别进行curvelet变换,得到各自的低频系数和高频系数。利用低频系数取平均,高频系数取区域能量值取大者的融合规则,得到融合图像的curvelet系数。最后进行curvelet逆变换,得到融合图像。结果得到了全幅清晰的融合图像。结论与传统的小波变换法和cur-velet变换法相比,所获得的融合图像更清晰,边缘处更平滑,与理想图像的差距更小。
Aim To combine the curvelet transform of the second generation with image regional energy to achieve the multi-focus image fusion. Methods Two multi-focus images are decomposed using curvelet transform, then the curvelet low-bands coefficients and the high-bands coefficients are acquired. The low-bands coefficients are used in the method of averaging, and the high-bands coefficients are used in the method of choosing the max of the regional energy. Finally, the fused image is obtained by the inverse eurvelet transform. Results A clear fused image is acquired. Conclusion It is shown by the experiment that clearer fused image and smoother edge can be obtained by this method, which has less difference on the ideal image, compared with conventional wavelet transform method and curvelet transform method.
出处
《西北大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第5期750-754,共5页
Journal of Northwest University（Natural Science Edition）
基金
陕西省自然科学基金资助项目(2005A12)
作者简介
郭敏，女，陕西师范大学教授，博士后，从事模式识别、图像处理、数据融合等方面的研究。
