摘要
针对多聚焦图像融合时间较长、融合效果较差的问题,提出一种分层模型的图像融合方法。首先,根据源图像模糊区域分布情况选取恰当的网格数以得出融合模型,并计算源图像所对应模块的灰度方差值,以决定区域子模块的层次划分,从而得到最佳融合区域;在此基础上,利用加权平均法解决不同模块边缘区域的不连续性,实现多聚焦图像的快速融合。最后,仿真结果验证了所提出方法的有效性与可行性。
The issue of hierarchical model for multi-focus image fusion was proposed to improve the visual effect and real-time performance. According to the ambiguous zones of source image, the appropriate grid value was selected and the fusion model was proposed. By using fuzzy logic method and threshold segmentation, the gray variance values of the corresponding block were calculated and the hierarchical classification of sub-block was given to achieve the best fusion zones. Based on these, the weighted mean method was proposed to smooth the discontinuity of different block marginal area. These guaranteed the real-time performance of image fusion. The simulation verifies the feasibility and effectiveness of the proposed theoretical results.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2016年第6期1372-1379,共8页
Journal of System Simulation
基金
国家自然科学基金(61203033)
安徽省自然科学基金(1608085MF146)
安徽省高校自然科学研究重点项目(KJ2016A062)
关键词
分层模型
多聚焦图像
快速融合
方差值
hierarchical model
multi-focus image
fast fusion
variance values
