摘要
无衍射激光图像一些小特征,容易产生干涉衍射偏移,导致很难提取,为了提高对无衍射激光图像微小监测能力,提出一种无衍射激光图像微小特征提取方法。通过提取激光图像微小特征的干涉稀疏散乱点,引入Hessian尺度分解方法进行小特征颜色分量RGB分解,利用匹配约束函数对RGB分量进行加速鲁棒特征匹配,以此实现图像微小特征提取;实验结果表明,采用该方法进行无衍射激光图像微小特征提取的准确性较好,输出图像的归一化相关系数值较高,说明该方法的抗干涉衍射干扰性较强。
Some tiny characteristics of non-diffraction laser image are easy to make diffraction shift which makes it a hard extraction. To improve the ability of monitoring tiny non-diffraction laser image, put forward a kind of tiny characteristics extraction method of non-diffraction laser image. Introduce Hessian scale decomposition method for small feature color components RGB through extracting tiny sparse scattered points of laser image tiny characteristics. Use matching constraint functions to accelerate the robustness feature matching of RGB components to achieve tiny fea- ture extraction. Experimental results show that this method has good accuracy, high related values of output image nor- malization which shows that this method has strong anti-interference ability.
出处
《激光杂志》
北大核心
2018年第3期155-158,共4页
Laser Journal
基金
莱芜市科技局自然科学基金资助项目(No.20140346)
关键词
无衍射激光图像
微小
特征
提取
加速鲁棒特征匹配
non-diffraction laser images
tiny
characteristic
extraction
accelerate the robust feature matching
