摘要
为了提高图像纹理特征提取的查准率，采用图论萤火虫搜索算法。首先基于图论分层结构，确定萤火虫分布；接着通过差分操作更新萤火虫，依据轮盘法运动、欧式距离计算萤火虫的相互之间吸引度；然后对萤火虫搜索成功与失败分析，确定动态移动步长；最后建立图像纹理特征提取模型和算法流程。实验显示该算法既能提高查准率，又缩短处理时间。
In order to improve the precision of image texture feature extraction, the graph theory glowworm search algorithm （GTGS） was proposed. Firstly, glowworm was distributed based on hierarchical architecture of graph theory. Secondly, glowworm was updated with differential operation, and moved by roulette wheel method, Euclidean distance was computed each attractiveness of glowworm. Thirdly, dynamic moving length was determined with success and failure of glowworm. Finally, image texture feature extraction model was established and algorithm was given. Simulation shows that GTGS algorithm precision is better, and the time is less.
出处
《计量学报》
CSCD
北大核心
2016年第3期255-259,共5页
Acta Metrologica Sinica
基金
河南省科技攻关项目(132102210523)
关键词
计量学
萤火虫算法
图论
分层结构
纹理特征
metrology
glowworm search algorithm
graph theory
hierarchical architecture
texture feature
