【摘要】 由于自身特有的链式群更新模式，基于原始樽海鞘群优化的图像匹配方法在一定程度上可降低陷入局部最优的概率，该方法在匹配速度、匹配时间以及匹配精度上仍有不足。因此，本文提出一种基于改进樽海鞘群优化的图像匹配方法。使用立方混沌初始化种群，调整收敛因子变化趋势，使种群尽可能遍历整个搜索空间，以此增强全局搜索能力；对跟随者进行正交方向的扰动，避免跟随者进行盲目的曲线搜索，以扩大其搜索范围；引入寻优者，致力于开发当前最优点附近的搜索空间，使算法加快搜索到全局最优点，提高算法速度。仿真结果表明，与基于粒子群优化(Particle swarm optimization, PSO)、蚁狮优化(Ant lion optimizer, ALO)和樽海鞘群优化(Salp swarm algorithm, SSA)3种算法的图像匹配方法相比，本算法提高了全局搜索能力，有效地降低了匹配时间，在收敛速度、收敛精度以及鲁棒性上有较好表现。更多还原

【Abstract】 Because of the unique chained update pattern of Salp swarm algorithm(SSA), the image matching based on salp swarm algorithm effectively reduces the probability of the population falling into local optimum. But there are some deficiencies in this algorithm, such as the low matching speed, much computing time and low matching accuracy. So an image matching based on improved salp swarm algorithm was proposed. The populations were initialized by cube mapping to ensure a uniform distribution of populations throughout the search space, and the trend of convergence factor was adjusted to enhance the global search capability of the algorithm. The followers were applied with perturbations in the orthogonal directions to avoid a large number of repeated searches. In order to speed up the search up the search to the global optimal point, the optimal point explorers were applied which work to develop the search space near the current optimal point. The simulating results confirmed that the proposed algorithm has higher global search capability and lower matching time than those of image matching based on Particle swarm optimization(PSO), Ant lion optimizer(ALO) and Salp swarm algorithm(SSA). Meanwhile, the proposed algorithm has better performance in terms of convergence speed, convergence accuracy and noise immunity compared to the other three algorithms.更多还原