基于滚动时域粒子群优化的视频去雾算法

无人机视频由于拍摄的位置和场景不断移动,环境参数亦不断变化,采用以往针对固定场景的去雾方法不能达到最佳效果.为了使无人机视频去雾算法具有自适应性,提出一种基于滚动时域粒子群优化的视频去雾算法.将基于周期和事件混合驱动的滚动调度策略与粒子群算法(PSO)相结合,对可调去雾参数进行滚动优化调整,当与上次优化的帧间隔数大于阈...

Receding horizon particle swarm optimization based video defogging algorithm

In order to solve the problems of high computational complexity, strong parameter dependence and weak global optimization ability of traditional swarm intelligence optimization algorithm, a fruit fly optimization algorithm based on double drive with the theory of bacterial chemotaxis is proposed. Considering the distribution of the superior and the inferior fruit fly groups, the concepts of multiple repellents and multiple attractants are proposed, and the location of fruit fly is updated under the double drive, so as to avoid the invalid search of the traditional methods which only depend on the local best (worst) position during the updating process. Then, based on the fitness value information of the fruit flies, a weighted centroid vector calculation method of multiple repellents and multiple attractants is proposed to determine the searching radiuses of fruit flies adaptively and avoid the problem of strong parameter dependence faced by traditional methods. The experimental results on standard functions show that, the proposed method has lower parameter dependence, higher convergence accuracy and convergence speed than existing typical algorithms. Moreover, the PID controller optimized by the proposed method has high response speed and stability, showing the ability of the proposed algorithm on PID parameter optimization.