图像融合与分裂合并相结合的分割算法研究

针对传统的分裂合并算法中存在的计算复杂度和分割后易于破坏边界信息的问题,提出一种将图像融合与分裂合并算法相结合的分割策略。首先对原始图像降低分辨率,抑制部分噪声,降低细节干扰,减少后期计算量;接着使用自适应边缘提取算法提取边缘信息并对优化图像进行分割;然后将处理后图像进行融合并还原尺度,以此来改善分裂合并算法带来的边界破坏问题。实验结果表明,该方法能够缓解传统算法中的运算量问题,尤其是边界破坏的问题,且分割效果较好。

Fusion combined with split-merge segmentation algorithm

An image segmentation algorithm of fusion combined with split￣merge is proposed.Firstly,lower the resolution of the original image to inhibit a part of the noise and reduce the interference of some details;Secondly,get the boundary information with adaptive￣edge algorithm and divide the optimization image;Finally,transform the fusion image to original scale.Experimental results show that,it could reduce the computational complexity in the traditional algorithm,and could prevent boundary information from being damaged partly.