Application of Distributed Genetic Algorithm Based on Migration Strategy in Image Segmentation

The traditional 1 dimension maximum between-class variance (1DMBV) method cannot obtain ideal threshold if the image has low SNR, while 2DMBV method can perform well even on the image with low SNR and low contrast, but with large computation. Some researchers combined the standard genetic algorithm with 2DMBV method (SGA-2DMBV), but it was premature and slowly convergent. In this paper, combined with 2DMBV, one distributed genetic algorithm base on migration strategy (DGA-2DMBV) was introduced to search optimal threshold with considerations to restraining premature convergence and shortening running time. Simulation results show that the proposed method is better than SGA-2DMBV at global search ability and far more quickly than 2DMBV at running time.