DSP For Image Segmentation Based On FODPSO Algorithm

Image segmentation has been mainly used for extraction of objection, and is the basis for image classification and pattern recognition. A new novel method called fractional-order Darwinian particle swarm algorithm is presented in this paper. The algorithm presents fractional calculus strategy to control the convergence of particles with memory effect and is able to determine the n-1 optimal for n-level threshold on a given image.This paper mainly researches image segmentation methods based on swarm intelligence optimization algorithm. Owing to different swarm intelligence algorithms’ adaptability, it is necessary to test different swarm intelligence algorithm performance, so that the optimal algorithm is chosen for image segmentation on DSP.Firstly, the genetic algorithm, simulated annealing algorithm, ant colony algorithm, seeker optimization algorithm, PSO, APSO, CFPSO, Darwinian particle swarm algorithm, fractional-order Darwinian particle swarm optimization algorithm were analyzed respecti-vely, focusing on algorithm theory and algorithm procedure in detail.Secondly, Owing to different algorithm adaptability, mean objective function value, standard deviation of objective function value and CPU time index were chosen to evaluate algorithm’s performance. Many experiments were designed to research on algorithm’s performance, such as double objective optimization design, three objective optimization design and multi-scale image segmentation. Compared with GA, SA, ACO, SOA, PSO, APSO, CFPSO, DPSO algorithm, testing results show that FODPSO algorithm enhanced performance for convergence speed, stability, solution accuracy and global optimality, and greatly overcame the shortcomings of traditional methods, such as local optima and slow convergence speed.Thirdly, TMS320VC5509A has the characteristics of small size, low cost and reliable performance, is good at image processing. In this experiment, test images were shot by OV7670 CMOS camera, and segmented by FODPSO algorithm based on CCS3.3 platform.At last. Algorithms involved in this paper were simulated by MATLAB codes, and then FODPSO algorithm was transplanted into DSP based on CCS3.3 platform to complete the multi-dimensional image segmentation experiments.DSP image segmentation research has important significance for practical application.