基于粒子群和蚁群算法的枪弹图像边缘检测方法

Bullet Image Edge Detection Based on Algorithm of Particle Swarm and Ant Colony

针对传统枪弹图像边缘检测中存在的收敛速度慢等缺陷,本文提出了一种基于粒子群和蚁群算法的枪弹图像边缘检测方法——群体优化算法.该算法的关键是把上述两种算法结合起来,使算法同时具有多样性和正反馈.首先将图像进行粒子群优化(PSO),在满足预设收敛条件后,将其次优解转换为蚁群优化(ACO)的初始信息素分布.然后,执行蚁群优化运算.当蚂蚁寻找食物时,多样性避免蚂蚁进入无限循环.当PSO达到预定收敛条件后,能够保持良好的正反馈.最后,显示图像边缘信息.实验结果表明,所提出的优化算法能够获取清晰连续的枪弹图像边缘信息,并且细节完整、搜索效率高.

Aiming at the defects such as slow convergence speed in the traditional edge detection of bullet images, a group optimization algorithm was proposed, which was an edge detection method for bullet image based on particle swarm optimization and ant colony optimization. The key was to combine the a- bove two algorithms together, so that the algorithm had both diversity and positive feedback. Firstly, PSO operation on the image was performed, and the sub-optimal solution of PSO was converted to the distribution of the initial pheromone for the following ACO after meeting the preset convergence condi- tions. Then ACO operation was performed. When ants were looking for food, diversity prevented the ants from going into an infinite loop. When the PSO reached a predetermined convergence condition, good positive feedback could be maintained. Finally, the bullet edge information of the image was shown. The experimental result proves that the hybrid algorithm proposed can extract clear bullet edges with complete details and profound and continuous edge information and that this algorithm works effec- tively in image edge detection.