Image Threshold Segmentation Based On OTSU Algorithm

Image segmentation is an important step in the process of image analysis.Among many image segmentation methods,threshold segmentation is a kind of widely used method in image segmentation.Its basic principle is to calculate the segmentation threshold based on image gray histogram,and then segment the image according to the threshold.In image segmentation,threshold segmentation is widely used.The main principle of threshold segmentation method is to calculate the information of image grayscale histogram to get the segmentation threshold.As the representative algorithm of threshold method,OTSU method has the advantages of simple principle,high segmentation accuracy and stable algorithm.The core idea of OTSU is to maximize the variance between classes to analyze and calculate the optimal threshold.In order to be able to process more complex images with OTSU method,a large number of scholars have extended it to two and three dimensions.Because of the increase of threshold number,the algorithm has a huge amount of computation and a significant decrease in efficiency.In this paper,the computational complexity and segmentation accuracy of the algorithm are considered.Through the analysis of 2D and 3D OTSU method,two improved methods are obtained,which can greatly improve the efficiency of the algorithm,and the segmentation effect is good.The feasibility of the improved method will be verified by experiments.Specific improvements can be summarized as follows:(1)Improved two-dimensional OTSU method for wolf swarm optimization.The threshold selection of the original two-dimensional OTSU algorithm generally depends on the exhaustive search method,which has a large amount of computation and poor real-time performance,which affects the efficiency of image segmentation.In order to reduce the running time of the algorithm,this paper analyzes and studies the wolf swarm algorithm and two-dimensional OTSU algorithm,and proposes an improved OTSU image segmentation method based on the improved wolf swarm optimization algorithm.The idea of solving the current local optimum in PSO algorithm is introduced into the wandering and calling behavior of the wolf swarm algorithm,so that the information exchange among the wolves can be realized,and the accuracy of searching the best threshold value can be improved.Adaptive besieging behavior is adopted to speed up the optimization of the algorithm.(2)The three-dimensional OTSU method is improved.In addition to the gray value of pixel points,the three-dimensional OTSU method introduces the gray mean value and the median gray value of two thresholds.Because the OTSU method uses exhaustive search strategy,the calculation of the 3D OTSU method needs three cycles,which leads to the huge amount of calculation of thealgorithm.In order to obtain the optimal threshold,the wolf swarm algorithm is introduced to obtain the optimal threshold by walking,calling and besieging these three intelligent behaviors and the information interaction among the wolves.This can speed up the search and reduce computing time.In order to avoid the improved algorithm falling into local optimum,chaotic optimization method was introduced after the besieged behavior of the wolves,and the chaotic optimization search was carried out on the sub-optimal solution of the algorithm.The chaotic optimization method is a search process that maps the computation process of the algorithm to a chaotic trajectory,which has the advantage of avoiding the algorithm from falling into local extremum.