Research On Image Segmentation Algorithm Using Variable Precision Rough Sets And Particle Swarm Optimization

Image processing is an important way to obtain information and has been widely applied to military,medical,and transportation and other important fields.Image segmentation plays an important role in image processing and can be studied in image analysis,image understanding,and image interpretation.At present,image segmentation is no universal segmentation algorithm.Threshold segmentation is one of the commonly used segmentation algorithms.It is simple,intuitive,and easy to implement.It is also a research and application hot spot.Considering the inherent complexity and correlation of the image,the uncertainty and fuzziness in the image segmentation process.How to handle the uncertainty in the image segmentation process so as to obtain more accurate segmentation results is a difficult point in image segmentation.In this paper,the uncertainty of image segmentation process is dealt with by the theory of variable precision rough and information entropy.The particle swarm optimization algorithm is used to improve the segmentation efficiency,and some threshold segmentation algorithms with good adaptability and flexibility are proposed.The main research results obtained are as follows:1.An image single-threshold segmentation algorithm with maximum precision and particle swarm optimization is proposed.Rough set is an effective mathematical tool to deal with uncertainty and fuzziness.The greater the dependence degree in variable precision rough set theory,the closer the relationship between attributes.The algorithm expresses the image with variable precision rough set,and uses the maximum dependence and the particle swarm algorithm to solve the optimal segmentation threshold.Through experimental comparison analysis,this single threshold segmentation algorithm is superior to the maximum average entropy method.2.An image single threshold segmentation algorithm with variable maximum precision rough entropy and particle swarm is proposed.The algorithm still expresses the image with variable precision rough set.The bigger the rough entropy is,the more uncertainty in the image is.The best segmentation threshold is solved with the maximum rough entropy,and the particle swarm optimization algorithm is used to improve the segmentation efficiency.Through experimental verification,this singlethreshold segmentation algorithm is superior to the maximum average entropy method.3.In the variable precision rough set theory,when the boundary of the variable precision rough set decreases,the rough entropy decreases,and the smaller the boundary,the more accurate the image represented by the variable precision rough set.The threshold is taken at the minimum rough entropy,so a new rough entropy is constructed and called the least-squares rough entropy.In this paper,we use the variable precision least squares rough entropy to solve the optimal segmentation threshold,and use the particle swarm optimization algorithm to improve the segmentation efficiency.Then we propose variable-precision least squares rough entropy and particle swarm image single threshold segmentation algorithm.Experimental verification shows that the single-threshold segmentation algorithm is superior to the maximum average entropy method.It also shows that variable-precision rough entropy is able to deal with the uncertainty of the image segmentation process.