Improvement of grayscale image 2D maximum entropy threshold segmentation method

To increase the segmentation speed and efficiency, traditional 2D maximum entropy threshold segmentation method is improved. The improved segmentation method is called PSO-SDAIVE algorithm. In this new algorithm, the 2D gray histogram is changed and forms the 2D D -value attribute gray histogram. When computing image entropy, the spatial gray information of pixels is included in computation. The improved entropy is called spatial different attribute information value entropy (SDAIVE). Otherwise, Particle Swarm Optimization (PSO) algorithm is used to solve maximum of SDAIVE. The corresponding solution of SDAIVE maximum is taken as optimal image segmentation threshold. In experiment, segment different grayscale image and testify the algorithm performance. Experimental results show that PSO-SDAIVE algorithm can quickly and accurately obtain segmentation threshold. Compare with other segmentation method, the cost time of solving optimal threshold is short. Otherwise, this algorithm can better segment noise image.