基于改进布谷鸟算法的图像配准和融合中的参数优化

发布时间：2018-06-26 19:13

  本文选题：布谷鸟算法 + 遗传因子　； 参考：《河北大学》2016年硕士论文

【摘要】：图像配准可以归结为寻求最佳空间变换的多参数优化问题,图像融合中加权系数等参数经过优化后会使融合效果更佳。快速、精确、适应性强的优化算法是实现参数优化的重要步骤。标准布谷鸟算法是基于布谷鸟寻窝产卵行为提出的新型智能算法,具有简单高效、随机路径优、参数少、程序运行简单等特征,但也存在局部搜索能力相对较弱、后期搜索速度慢、计算精度不高等缺点。将遗传因子融入标准布谷鸟算法中,提出一种基于遗传因子的布谷鸟算法,通过选择操作和添加交叉、变异因子,增加种群的多样性,提高算法全局搜索能力。实验表明,改进后算法的寻优精度比标准布谷鸟算法高,表现出更好的收敛性和稳定性。将混沌搜索融入标准布谷鸟算法,提出一种基于混沌搜索的布谷鸟算法,利用混沌运动随机性、遍历性的特点,使种群均匀分布,并增强算法跳出局部极值的能力。实验表明,改进后的算法在搜索空间小时搜索能力增强,但其收敛性和稳定性不如基于遗传因子的布谷鸟算法好。为了进一步提高算法的收敛速度和计算精度,将算法中的固定参数改为随迭代过程自适应变化的动态参数,在基于遗传因子布谷鸟算法的基础上提出基于遗传因子自适应布谷鸟算法。实验表明,基于遗传因子自适应布谷鸟算法的收敛速度和全局寻优能力进一步提高并且可靠性更高。最后将基于遗传因子自适应布谷鸟算法用于图像配准和图像融合的参数优化中,并与其他智能算法对比,实验表明,该算法配准精度高、时间短,并且寻找的最优权值系数融合后得到的融合图像能提取更多有用信息,融合效果更佳,充分验证了本文算法的有效性、稳定性和可行性。
[Abstract]:Image registration can be attributed to the multi-parameter optimization problem of seeking the best spatial transformation. After the optimization of the parameters such as weighting coefficient in image fusion, the fusion effect will be better. Fast, accurate and adaptable optimization algorithm is an important step to realize parameter optimization. Standard Cuckoo algorithm is a new intelligent algorithm based on cuckoo nest and spawning behavior. It has the characteristics of simple and efficient, random path optimization, few parameters, simple program operation, etc. However, the local search ability is relatively weak. Late search speed is slow, calculation accuracy is not high shortcomings. The genetic factor is incorporated into the standard cuckoo algorithm, and a genetic factor-based cuckoo algorithm is proposed. By selecting the operation and adding the crossover and mutation factor, the diversity of the population and the global search ability of the algorithm are improved. Experimental results show that the improved algorithm has better convergence and stability than the standard Cuckoo algorithm. The chaotic search is incorporated into the standard cuckoo algorithm, and a chaotic search based cuckoo algorithm is proposed. The chaotic motion randomness and ergodicity are used to make the population distribute evenly, and the ability of the algorithm to jump out of the local extremum is enhanced. The experimental results show that the improved algorithm has a better ability to search in hours of searching space, but its convergence and stability are not as good as the genetic factor-based cuckoo algorithm. In order to further improve the convergence speed and accuracy of the algorithm, the fixed parameters in the algorithm are changed to the dynamic parameters that change adaptively with the iterative process. On the basis of genetic factor-based cuckoo algorithm, an adaptive genetic factor-based cuckoo algorithm is proposed. Experiments show that the convergence speed and global optimization ability of the adaptive cuckoo algorithm based on genetic factor are further improved and the reliability is higher. Finally, the genetic factor-based adaptive cuckoo algorithm is applied to the parameter optimization of image registration and image fusion, and compared with other intelligent algorithms, the experimental results show that the algorithm has high registration accuracy and short time. And the fusion image obtained by the fusion of the optimal weight coefficients can extract more useful information and the fusion effect is better, which fully verifies the effectiveness, stability and feasibility of the proposed algorithm.
【学位授予单位】：河北大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18;TP391.41
，

本文编号：2071113