摘要
基于双CPU的多车道交通流实时动态信息检测系统图像分割的阈值自动优化选取系统,通过将模拟退火思想引入到遗传算法中设计了退火遗传算法(AGA),以最大类方差函数作为遗传算法中适应度的评价函数,利用退火算法后期寻优能力强和遗传算法全局搜索能力强的特点,实现图像阈值的自动优化选取.MATLAB仿真数据表明,本算法较基本遗传算法寻优性能更强.将其应用于该交通流检测系统,增强了整个系统的实时性和鲁棒性.
In view of author＇s design of the real-time dynamic information examination system of multiple traffic flow＇s image division＇s question of threshold automatic optimizing selection based on double CPU, here we solved this problem through introducing the thoughts of simulation of annealing in the basic genetic algorithm to propose the annealing genetic algorithm（AGA）, with the Otsu method being used as the fitness evaluation fuction, using the characteristic of annealing algorithm＇s strong ability of partial searching and genetic algorithm＇s strong ability of overall searching. The MATLAB simulation data indicates that this algorithm has stronger optimal ability and faster convergence speed than the basic genetic algorithm. Appling it in this detecting system of traffic flow, the overall system＇s timeliness and robustness are strengthened.
出处
《微电子学与计算机》
CSCD
北大核心
2008年第8期22-24,28,共4页
Microelectronics & Computer
基金
国家自然科学基金项目(60664001)
江西省自然科学基金项目(0511030)
关键词
退火遗传算法
图像分割
交通流
实时性
annealing genetic algorithm
image division
traffic flow
timeliness
