Chaos Optimization Algorithm And Its Application In Traffic Image Segmentation

In the field of transportation,there are many types of traffic images,including license plate images,vehicle images,road images and traffic sign images.Since traffic images are sensitive to noise,it is of great significance to study the processing technology for traffic images.Segmentation results of traffic image directly affect its subsequent detection and recognition.Traditional clustering-based segmentation algorithms for traffic images have poor anti-noise ability and fall into local optimum easily.As the chaos optimization algorithm has strong global search capability,chaos optimization algorithm is studied in this paper and the weakness of chaos optimization algorithm is overcame.The improved parallel chaos harmony search algorithm is applied to the segmentation of traffic images and the clustering algorithm for traffic image segmentation based on parallel chaos harmony search is proposed.With chaos being applied to multi-objective optimization problems,multi-objective chaos optimization algorithm is proposed.In addition,the clustering algorithm for traffic image segmentation based on multi-objective chaos optimization is proposed.The main research work of this paper can be divided into the following four parts:Although the chaos optimization algorithm has strong global search ability,it is sensitive to initial value and its local search ability is weak.The parallel chaos search mechanism is introduced into chaos optimization algorithm to overcome its weakness of being sensitive to initial value.To improve the local search ability of chaos optimization algorithm,chaos search algorithm is combined with harmony search algorithm.With the introduction of crossover operation and merging operation,chaos is applied to multi-objective optimization problems and multi-objective chaos optimization algorithm is proposed.In the multi-objective chaos optimization algorithm,chaotic carrier wave is used to search globally,moreover,crossover operation and merging operation are used to search locally.In addition,self-adaptive noise is introduced into multi-objective chaos optimization algorithm to strengthen its adaptiveness and to improve its local search capability.A segmentation algorithm for traffic image is proposed by combining parallel chaos harmony search algorithm with fuzzy clustering technology.When compared to the hard clustering algorithm and fuzzy clustering algorithm,simulation and experimental results show that the proposed algorithm behaves much better.Another segmentation algorithm for traffic image is proposed by combining multi-objective chaos optimization algorithm with clustering technique.The clustering optimization indexes for the proposed algorithm are determined to be XB and Jm.In order to verify the effectiveness of the proposed algorithm,it is compared to the clustering algorithm based on parallel chaos harmony search and the clustering algorithm based on multi-objective genetic search.The simulation results show that the proposed algorithm does well in segmentation of traffic images.