摘要
目的提出一种空中目标特征识别的新算法。方法在轮廓线提取过程中,采用二维Bubble小波多分辨率分析来降噪并弱化伪边缘,采用GVFS法跟踪和提取目标体的边界,再利用多边形近似算法按指定的顶点数提取多边形,以使其和目标特征模板进行匹配。结果提出了一种六步算法。结论对直升机的空中图像进行了特征提取仿真,该方法对空中目标识别很有效。
Aim To propose a new algorithm for aerial target features identification. Methods In contour extraction, two dimensional Bubble wavelet is used to remove noises and weaken false edges. GVFS is used to track and extract accurate body boundary. Polygon approximation is used to extract a polygon with pre-specified vertex number so it can be used to match the target feature template. Results A six step algorithm of aerial target features identification using MRA-GVFS is proposed. Conclusion Algorithm for features identification simulation is applied to aerial target images of a helicopter. Algorithm is available.
出处
《西北大学学报（自然科学版）》
CAS
CSCD
北大核心
2006年第3期374-376,共3页
Journal of Northwest University（Natural Science Edition）
基金
国防预研基金资助项目(41327030103)
