当前位置:主页 > 科技论文 > 软件论文 >

结合分层处理的激光光条亚像素中心提取方法

发布时间:2018-06-10 19:12

  本文选题:激光光条提取 + 视觉测量 ; 参考:《红外与激光工程》2017年10期


【摘要】:为实现大型航空零件三维测量过程中激光光条的快速高精度提取,提出了一种结合分层处理的激光光条亚像素中心提取方法。首先,根据序列图像的结构不变性将高分辨率图像压缩为低分辨图像。接着,通过二次拟合求解低分辨率图像中激光光条中心的法线斜率。然后,将低分辨图像求得的法线斜率还原到高分辨激光光条图像中。进而通过灰度重心判断准则,快速计算激光光条的亚像素中心。最后,采用所提出的方法分别在实验室和大型航空零件装配测试台上进行了复合材料标准样件和复杂零件的三维形貌测量。实验结果表明:该方法的单激光光条重建误差为0.269 mm,三维形面的重建误差为0.268 mm。该方法可有效提高工程零件快速测量过程中激光光条提取精度,满足大型航空零件现场测量的工程要求。
[Abstract]:In order to realize the fast and high precision extraction of laser stripes in 3D measurement of large aerial parts, a method of subpixel center extraction of laser stripes combined with delamination is proposed. Firstly, the high resolution image is compressed to low resolution image according to the structure invariance of the sequence image. Then, the normal slope of the center of the laser strip in the low resolution image is solved by quadratic fitting. Then, the normal slope obtained from the low resolution image is reduced to the high resolution laser strip image. Furthermore, the sub-pixel center of laser beam is calculated quickly by the criterion of gray center of gravity. Finally, the 3D morphologies of composite standard samples and complex parts were measured in laboratory and large aerial parts assembly test bench using the proposed method. The experimental results show that the reconstruction error of single laser strip is 0.269 mm and that of 3D shape plane is 0.268 mm. This method can effectively improve the precision of laser light strip extraction in the process of rapid measurement of engineering parts, and meet the engineering requirements of field measurement of large aerial parts.
【作者单位】: 大连理工大学机械工程学院;
【基金】:国家973计划(2014CB46504) 国家自然科学基金(51375075,51227004);国家自然科学基金创新研究群体项目(51621064) 辽宁省自然科学基金(2014028010)
【分类号】:TP391.41;V261

【相似文献】

相关期刊论文 前2条

1 于起峰,孙祥一,权铁汉,雷志辉;用标定和亚像素技术实现三维运动目标的高精度测量[J];宇航学报;1999年03期

2 史少龙;尹达一;;改进型灰度质心实时算法研究[J];光电工程;2013年12期



本文编号:2004299

资料下载
论文发表

本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2004299.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户3b833***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com