Research On Artifical Fish Swarm Algorithm And Its Application In Image Enhancement

Optimization method is significant for solving life problems, and it is always the study topic of many experts and scholars. Compared with traditional evolutionary algorithm and gradient-based optimization algorithm, swarm intelligence optimization algorithm has many advantages such as good self-organization, simple individual behavior and strong global optimization capability. In recent years, swarm intelligence has become a hot research field. Artificial Fish Swarm Algorithm is an intelligent optimization algorithm which simulates fish’s foraging behavior, swarming behavior and following behavior. AFSA has advantages such as strong robust, good global convergence and low requirement for initial values, so it can be used to solve nonlinear function optimization problems effectively. But for some complex problems, AFSA still has some problems such as low optimization accuracy, long time consumed in each iteration and low convergence in the late optimization.To solve the above problem, this paper proposes an improved artificial fish swarm algorithm using the Various Population Strategy, and VAFSA for short. First, the improved algorithm adjusts moving step dynamically to increase optimization accuracy. Second, behavior choice with memory feature is proposed. When artificial fish chooses behavior, it simulates executing behavior in last iteration; if its result is better than the result of last behavior, the fish actually executes this behavior instead of simulating other all behaviors; this improvement reduces the time consumed each iteration. Third, various population strategy is introduced to increase the diversity in late evolution, and escape from local optimization faster. Experimental results show that VAFSA performs much better than AFSA that VAFSA improves optimization accuracy effectively, reduces the time consumed each iteration and accelerates the convergence in the late optimization.The basic principles of image enhancement are studied in this paper. Normalized non-complete Beta function is used to fit gray scale transformation function, mean square error is taken as the objective function and image enhancement problem is transformed into the problem of searching for optimal result of the objective function. A new image enhancement algorithm is proposed by using VAFSA. The experimental results show that images enhanced by this new image enhancement algorithm perform better in visual effects and image quality.How to select and adjust the parameters, how to develop Various Population Strategy and how to select behavior in iterations are the key factors affecting the performance. In the subsequent studies, the Various Population Strategy will be studied further to improve the performance of VASFA. In addition, application research of VAFSA in other research fields should be considered.