Image segmentation algorithm based on improved genetic algorithms and grey relational degree analysis

In order to obtain the optimal segmentation threshold and get rid of the local optimal solution of image segmentation, this paper reconstructs crossover and mutation rate which will not be zero at any time. Meanwhile, crossover and mutation genetic operations are used to search the optimal segmentation threshold, where the fitness function is the largest two-dimensional entropy function. Then, grey correlation analysis, which can get comprehensive correlation between regions, is performed on splitting images to guide region merging. Simulation results show that this method can effectively segment the target area with certain noise immunity.