摘要
阈值法分割图像时只利用图像的灰度信息,具有直观、实现简单的特点。针对传统的粒子群优化算法(Particle Swarm Optimization,PSO)分割图像易陷入局部最优的缺点,提出一种基于改进粒子群优化算法的Otsu图像阈值分割方法。以Otsu算法的类间方差作为适应度函数,在每次迭代中选取适应度较好的粒子同时加入新的粒子,以提高粒子多样性。实验表明,与Otsu算法和PSO算法相比,改进的粒子群优化算法不仅加快了收敛速度和运算速度,而且提高了图像分割的准确率。
The thresholding method only needs the gray information to spilt image,which is more intuitive and much easier to be implemented.Aiming at the problem that the traditional PSO algorithm used for image segmentation is easy to fall into local optimum,this paper proposed an Otsu image threshold segmentation method based on the improved PSO.We took the inter-class variance of Otsu as the fitness function,and selected the particles with better fitness and added new particles to increase the diversity of the particles.The experimental results show that,compared with Otsu methods and PSO algorithm,the improved PSO accelerates the speed of convergence and computation,and improves the accuracy of image segmentation.
出处
《计算机科学》
CSCD
北大核心
2016年第3期309-312,共4页
Computer Science
基金
青年科学基金项目(61402212)语义Web模糊规则互换与推理关键技术研究资助
关键词
图像分割
OTSU
类间方差
粒子群优化
适应度函数
Image segmentation
Otsu
Inter-class variance
Particle swarm optimization
Fitness function
